ci: copy ui-svelte/.npmrc before npm ci in fork-cuda build

npm ci ran without .npmrc (legacy-peer-deps=true), failing on the
tailwind/vite peer dependency conflict. Copy .npmrc with the manifest.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

.coderabbit.yaml

@@ -15,6 +15,8 @@ reviews:
   auto_review:
     enabled: false
     drafts: false
+  unit_tests:
+    enabled: false
 chat:
   auto_reply: true
 issue_enrichment:

.gitea/workflows/build-cuda-image.yml

@@ -0,0 +1,76 @@
+name: Build CUDA image (fork)
+
+# Builds this fork's llama-swap (serial scheduler + embedded UI) from source and
+# layers it on a pinned llama.cpp CUDA server base, then pushes to the Gitea
+# container registry, e.g. gitea.stevedudenhoeffer.com/steve/llama-swap:v230-cuda-b9821
+#
+# Requires repo secrets: REGISTRY_USER, REGISTRY_PASSWORD (push to the registry).
+
+on:
+  workflow_dispatch:
+    inputs:
+      llama_swap_version:
+        description: "llama-swap version label (image tag prefix)"
+        required: false
+        default: "v230"
+      llamacpp_build:
+        description: "llama.cpp CUDA server build (base image tag suffix)"
+        required: false
+        default: "b9821"
+  # Building the build definition itself kicks off a fresh image.
+  push:
+    branches: [main]
+    paths:
+      - ".gitea/workflows/build-cuda-image.yml"
+      - "docker/fork-cuda.Containerfile"
+
+env:
+  REGISTRY: gitea.stevedudenhoeffer.com
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Compute image metadata
+        id: meta
+        run: |
+          LS_VER="${{ inputs.llama_swap_version || 'v230' }}"
+          LCPP="${{ inputs.llamacpp_build || 'b9821' }}"
+          {
+            echo "image=${REGISTRY}/${{ github.repository }}"
+            echo "tag=${LS_VER}-cuda-${LCPP}"
+            echo "base_tag=server-cuda-${LCPP}"
+            echo "ls_version=${LS_VER}"
+            echo "build_date=$(date -u +%Y-%m-%dT%H:%M:%SZ)"
+          } >> "$GITHUB_OUTPUT"
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Log in to Gitea registry
+        uses: docker/login-action@v3
+        with:
+          registry: ${{ env.REGISTRY }}
+          username: ${{ secrets.REGISTRY_USER }}
+          password: ${{ secrets.REGISTRY_PASSWORD }}
+
+      - name: Build and push
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          file: docker/fork-cuda.Containerfile
+          push: true
+          provenance: false
+          build-args: |
+            BASE_TAG=${{ steps.meta.outputs.base_tag }}
+            LS_VERSION=${{ steps.meta.outputs.ls_version }}
+            GIT_HASH=${{ github.sha }}
+            BUILD_DATE=${{ steps.meta.outputs.build_date }}
+          tags: ${{ steps.meta.outputs.image }}:${{ steps.meta.outputs.tag }}
+
+      - name: Summary
+        run: |
+          echo "Pushed ${{ steps.meta.outputs.image }}:${{ steps.meta.outputs.tag }}" >> "$GITHUB_STEP_SUMMARY"

.github/workflows/config-schema.yml

@@ -44,13 +44,10 @@ jobs:
 
           echo "✓ config-schema.json is valid"
 
-      - name: Set up Python
-        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 #v6.2.0
+      - name: Set up Go
+        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c #6.4.0
         with:
-          python-version: "3.x"
-
-      - name: Install check-jsonschema
-        run: pip install check-jsonschema
+          go-version-file: go.mod
 
       - name: Validate config.example.yaml against schema
-        run: check-jsonschema --schemafile config-schema.json config.example.yaml
+        run: go test ./internal/config/ -run TestConfig_ExampleMatchesSchema -v

.github/workflows/containers.yml

@@ -2,10 +2,10 @@ name: Build Containers
 
 on:
   # time has no specific meaning, trying to time it after
-  # the llama.cpp daily packages are published
+  # the llama.cpp daily packages have time to build and publish (~8hr after llama.cpp project's cron)
   # https://github.com/ggml-org/llama.cpp/blob/master/.github/workflows/docker.yml
   schedule:
-    - cron: "37 5 * * *"
+    - cron: "00 12,18 * * *"
 
   # Allows manual triggering of the workflow
   workflow_dispatch:

.github/workflows/release.yml

@@ -32,11 +32,9 @@ jobs:
         uses: actions/setup-node@48b55a011bda9f5d6aeb4c2d9c7362e8dae4041e # 6.4.0
         with:
           node-version: "24"
-      - name: Install dependencies and build UI
+      - name: Build UI
         run: |
-          cd ui-svelte
-          npm ci
-          npm run build
+          make ui
 
       - name: Run GoReleaser
         uses: goreleaser/goreleaser-action@1a80836c5c9d9e5755a25cb59ec6f45a3b5f41a8 #7.2.1

.gitignore

@@ -8,4 +8,3 @@ dist/
 
 # UI build output; placeholder.txt is kept so the go:embed succeeds.
 internal/server/ui_dist/*
-!internal/server/ui_dist/placeholder.txt

AGENTS.md

@@ -5,16 +5,14 @@ llama-swap is a light weight, transparent proxy server that provides automatic m
 ## Tech stack
 
 - golang
-- typescript, vite and svelt5 for UI (located in ui/)
+- typescript, vite and svelte 5 for UI (located in ui-svelte/)
 
 ## Workflow Tasks
 
 - when summarizing changes only include details that require further action
-- just say "Done." when there is no further action
-- use the github CLI `gh` to create pull requests and work with github
 - Rules for creating pull requests:
-  - keep them short and focused on changes.
-  - never include a test plan
+  - keep them short and focused on changes
+  - skip the test plan
   - write the summary using the same style rules as commit message
 
 ## Testing
@@ -30,7 +28,7 @@ llama-swap is a light weight, transparent proxy server that provides automatic m
 ### Commit message example format:
 
 ```
-proxy: add new feature
+internal/server: add new feature
 
 Add new feature that implements functionality X and Y.
 

Makefile

@@ -19,21 +19,17 @@ all: mac linux simple-responder
 clean:
 	rm -rf $(BUILD_DIR)
 
-proxy/ui_dist/placeholder.txt:
-	mkdir -p proxy/ui_dist
-	touch $@
-
 # use cached test results while developing
-test-dev: proxy/ui_dist/placeholder.txt
-	go test -short ./proxy/... ./internal/...
-	staticcheck ./proxy/... ./internal/... || true
+test-dev:
+	go test -short ./...
+	staticcheck ./... || true
 
-test: proxy/ui_dist/placeholder.txt
-	go test -short -count=1 ./proxy/... ./internal/...
+test:
+	go test -short -count=1 ./internal/...
 
 # for CI - full test (takes longer)
-test-all: proxy/ui_dist/placeholder.txt
-	go test -race -count=1 ./proxy/... ./internal/...
+test-all:
+	go test -race -count=1 ./internal/...
 
 ui/node_modules:
 	cd ui-svelte && npm install
@@ -41,8 +37,7 @@ ui/node_modules:
 # build react UI
 ui: ui/node_modules
 	cd ui-svelte && npm run build
-	mkdir -p internal/server/ui_dist
-	cp -R proxy/ui_dist/. internal/server/ui_dist/
+	touch internal/server/ui_dist/placeholder.txt
 
 # Build OSX binary
 mac: ui
@@ -65,7 +60,7 @@ windows: ui
 	@echo "Building Windows binary..."
 	GOOS=windows GOARCH=amd64 go build -ldflags="-X main.commit=${GIT_HASH} -X main.version=local_${GIT_HASH} -X main.date=${BUILD_DATE}" -o $(BUILD_DIR)/$(APP_NAME)-windows-amd64.exe
 
-# for testing proxy.Process
+# for testing with real external processes
 simple-responder:
 	@echo "Building simple responder"
 	GOOS=darwin GOARCH=arm64 go build -o $(BUILD_DIR)/simple-responder_darwin_arm64 cmd/simple-responder/simple-responder.go

README.md

@@ -88,10 +88,11 @@ Real time log streaming:
 llama-swap can be installed in multiple ways
 
 1. Docker
-2. Homebrew (OSX and Linux)
-3. WinGet
-4. From release binaries
-5. From source
+2. Homebrew (macOS and Linux)
+3. MacPorts (macOS)
+4. WinGet
+5. From release binaries
+6. From source
 
 ### Docker Install ([download images](https://github.com/mostlygeek/llama-swap/pkgs/container/llama-swap))
 
@@ -155,6 +156,16 @@ brew install llama-swap
 llama-swap --config path/to/config.yaml --listen localhost:8080
 ```
 
+### MacPorts (macOS)
+
+> [!NOTE]
+> Maintained by MacPorts community - [llama-swap port](https://ports.macports.org/port/llama-swap). It is not an official part of llama-swap.
+
+```shell
+sudo port install llama-swap
+llama-swap --config path/to/config.yaml --listen localhost:8080
+```
+
 ### WinGet Install (Windows)
 
 > [!NOTE]

cmd/legacy/llama-swap.go

@@ -1,249 +0,0 @@
-package main
-
-import (
-	"context"
-	"flag"
-	"fmt"
-	"net/http"
-	"os"
-	"os/signal"
-	"path/filepath"
-	"strings"
-	"sync"
-	"syscall"
-	"time"
-
-	"github.com/gin-gonic/gin"
-	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/mostlygeek/llama-swap/internal/event"
-	"github.com/mostlygeek/llama-swap/internal/logmon"
-	"github.com/mostlygeek/llama-swap/internal/perf"
-	"github.com/mostlygeek/llama-swap/internal/watcher"
-	"github.com/mostlygeek/llama-swap/proxy"
-)
-
-var (
-	version string = "0"
-	commit  string = "abcd1234"
-	date    string = "unknown"
-)
-
-func main() {
-	// Define a command-line flag for the port
-	configPath := flag.String("config", "config.yaml", "config file name")
-	listenStr := flag.String("listen", "", "listen ip/port")
-	certFile := flag.String("tls-cert-file", "", "TLS certificate file")
-	keyFile := flag.String("tls-key-file", "", "TLS key file")
-	showVersion := flag.Bool("version", false, "show version of build")
-	watchConfig := flag.Bool("watch-config", false, "Automatically reload config file on change")
-	mainLogger := logmon.New()
-
-	flag.Parse() // Parse the command-line flags
-
-	if *showVersion {
-		fmt.Printf("version: %s (%s), built at %s", version, commit, date)
-		os.Exit(0)
-	}
-
-	conf, err := config.LoadConfig(*configPath)
-	if err != nil {
-		mainLogger.Errorf("Error loading config: %v", err)
-		os.Exit(1)
-	}
-
-	if len(conf.Profiles) > 0 {
-		mainLogger.Warn("Profile functionality has been removed in favor of Groups. See the README for more information.")
-	}
-
-	switch strings.ToLower(strings.TrimSpace(conf.LogLevel)) {
-	case "debug":
-		mainLogger.SetLogLevel(logmon.LevelDebug)
-	case "info":
-		mainLogger.SetLogLevel(logmon.LevelInfo)
-	case "warn":
-		mainLogger.SetLogLevel(logmon.LevelWarn)
-	case "error":
-		mainLogger.SetLogLevel(logmon.LevelError)
-	default:
-		mainLogger.SetLogLevel(logmon.LevelInfo)
-	}
-
-	mainLogger.Debugf("PID: %d", os.Getpid())
-
-	if mode := os.Getenv("GIN_MODE"); mode != "" {
-		gin.SetMode(mode)
-	} else {
-		gin.SetMode(gin.ReleaseMode)
-	}
-
-	// Validate TLS flags.
-	var useTLS = (*certFile != "" && *keyFile != "")
-	if (*certFile != "" && *keyFile == "") ||
-		(*certFile == "" && *keyFile != "") {
-		fmt.Println("Error: Both --tls-cert-file and --tls-key-file must be provided for TLS.")
-		os.Exit(1)
-	}
-
-	// Set default ports.
-	if *listenStr == "" {
-		defaultPort := ":8080"
-		if useTLS {
-			defaultPort = ":8443"
-		}
-		listenStr = &defaultPort
-	}
-
-	var mon *perf.Monitor
-	if !conf.Performance.Disabled {
-		mon, err = perf.New(conf.Performance, mainLogger)
-		if err != nil {
-			mainLogger.Errorf("failed to create monitor: %s", err.Error())
-			os.Exit(1)
-		}
-		mon.Start()
-	} else {
-		mainLogger.Info("performance monitoring is disabled")
-	}
-
-	// Setup channels for server management
-	exitChan := make(chan struct{})
-	sigChan := make(chan os.Signal, 1)
-	signal.Notify(sigChan, syscall.SIGINT, syscall.SIGTERM, syscall.SIGHUP)
-
-	// Context that bounds the lifetime of background watcher goroutines.
-	watcherCtx, watcherCancel := context.WithCancel(context.Background())
-
-	// Create server with initial handlergit
-	srv := &http.Server{
-		Addr: *listenStr,
-	}
-
-	// Support for watching config and reloading when it changes
-	reloading := false
-	var reloadMutex sync.Mutex
-	reloadProxyManager := func() {
-		reloadMutex.Lock()
-		if reloading {
-			reloadMutex.Unlock()
-			return
-		}
-		reloading = true
-		reloadMutex.Unlock()
-		defer func() {
-			reloadMutex.Lock()
-			reloading = false
-			reloadMutex.Unlock()
-		}()
-
-		if currentPM, ok := srv.Handler.(*proxy.ProxyManager); ok {
-			mainLogger.Info("Reloading Configuration")
-			conf, err = config.LoadConfig(*configPath)
-			if err != nil {
-				mainLogger.Warnf("Unable to reload configuration: %v", err)
-				return
-			}
-
-			mainLogger.Debug("Configuration Changed")
-			currentPM.Shutdown()
-			if mon != nil {
-				mon.UpdateConfig(conf.Performance)
-			}
-			newPM := proxy.New(conf)
-			newPM.SetVersion(date, commit, version)
-			newPM.SetPerfMonitor(mon)
-			srv.Handler = newPM
-			mainLogger.Debug("Configuration Reloaded")
-
-			// wait a few seconds and tell any UI to reload
-			time.AfterFunc(3*time.Second, func() {
-				event.Emit(proxy.ConfigFileChangedEvent{
-					ReloadingState: proxy.ReloadingStateEnd,
-				})
-			})
-		} else {
-			conf, err = config.LoadConfig(*configPath)
-			if err != nil {
-				mainLogger.Errorf("Unable to load configuration: %v", err)
-				os.Exit(1)
-			}
-			newPM := proxy.New(conf)
-			newPM.SetVersion(date, commit, version)
-			newPM.SetPerfMonitor(mon)
-			srv.Handler = newPM
-		}
-	}
-
-	// load the initial proxy manager
-	reloadProxyManager()
-
-	if *watchConfig {
-		go func() {
-			absConfigPath, err := filepath.Abs(*configPath)
-			if err != nil {
-				mainLogger.Errorf("watch-config unable to determine absolute path for watching config file: %v", err)
-				return
-			}
-			mainLogger.Info("Watching configuration for changes (poll-based, 2s interval)")
-			(&configwatcher.Watcher{
-				Path:     absConfigPath,
-				Interval: configwatcher.DefaultInterval,
-				OnChange: func() {
-					reloadProxyManager()
-				},
-			}).Run(watcherCtx)
-		}()
-	}
-
-	// Signal handling
-	go func() {
-		for {
-			sig := <-sigChan
-			switch sig {
-			case syscall.SIGHUP:
-				mainLogger.Debug("Received SIGHUP")
-				reloadProxyManager()
-			case syscall.SIGINT, syscall.SIGTERM:
-				mainLogger.Debugf("Received signal %v, shutting down...", sig)
-				if mon != nil {
-					mon.Stop()
-				}
-				watcherCancel()
-				ctx, cancel := context.WithTimeout(context.Background(), time.Second*5)
-				defer cancel()
-
-				if pm, ok := srv.Handler.(*proxy.ProxyManager); ok {
-					pm.Shutdown()
-				} else {
-					mainLogger.Error("srv.Handler is not of type *proxy.ProxyManager")
-				}
-
-				if err := srv.Shutdown(ctx); err != nil {
-					mainLogger.Errorf("Server shutdown: %v", err)
-				}
-				close(exitChan)
-				return
-			default:
-				// do nothing on other signals
-			}
-		}
-	}()
-
-	// Start server
-	go func() {
-		var err error
-		if useTLS {
-			mainLogger.Infof("llama-swap listening with TLS on https://%s", *listenStr)
-			err = srv.ListenAndServeTLS(*certFile, *keyFile)
-		} else {
-			mainLogger.Infof("llama-swap listening on http://%s", *listenStr)
-			err = srv.ListenAndServe()
-		}
-		if err != nil && err != http.ErrServerClosed {
-			mainLogger.Errorf("Fatal server error: %v", err)
-			os.Exit(1)
-		}
-	}()
-
-	// Wait for exit signal
-	<-exitChan
-}

config-schema.json

@@ -82,6 +82,78 @@
             },
             "additionalProperties": false,
             "description": "Timeout settings for proxy connections."
+        },
+        "groupsConfig": {
+            "type": "object",
+            "additionalProperties": {
+                "type": "object",
+                "required": [
+                    "members"
+                ],
+                "properties": {
+                    "swap": {
+                        "type": "boolean",
+                        "default": true,
+                        "description": "Controls model swapping behaviour within the group. True: only one model runs at a time. False: all models can run together."
+                    },
+                    "exclusive": {
+                        "type": "boolean",
+                        "default": true,
+                        "description": "Controls how the group affects other groups. True: causes all other groups to unload when this group runs a model. False: does not affect other groups."
+                    },
+                    "persistent": {
+                        "type": "boolean",
+                        "default": false,
+                        "description": "Prevents other groups from unloading the models in this group. Does not affect individual model behaviour."
+                    },
+                    "members": {
+                        "type": "array",
+                        "items": {
+                            "type": "string"
+                        },
+                        "description": "Array of model IDs that are members of this group. Model IDs must be defined in models."
+                    }
+                }
+            },
+            "description": "A dictionary of group settings. Provides advanced controls over model swapping behaviour. Model IDs must be defined in models. A model can only be a member of one group. Behaviour controlled via swap, exclusive, persistent."
+        },
+        "matrixConfig": {
+            "type": "object",
+            "description": "Solver-based alternative to groups. Declares valid combinations of concurrent models. The solver minimizes eviction cost when swapping. A config must use either groups or matrix, not both.",
+            "required": [
+                "vars",
+                "sets"
+            ],
+            "properties": {
+                "vars": {
+                    "type": "object",
+                    "description": "Short names for models. Keys must be alphanumeric, 1-8 characters. All sets and evict_costs must use these IDs.",
+                    "minProperties": 1,
+                    "additionalProperties": {
+                        "type": "string"
+                    },
+                    "propertyNames": {
+                        "pattern": "^[a-zA-Z0-9]{1,8}$"
+                    }
+                },
+                "evict_costs": {
+                    "type": "object",
+                    "description": "Relative cost of evicting a running model. Models not listed default to 1. Values must be positive integers.",
+                    "additionalProperties": {
+                        "type": "integer",
+                        "minimum": 1
+                    }
+                },
+                "sets": {
+                    "type": "object",
+                    "description": "Named sets of concurrent model combinations. Values are DSL strings using & (AND), | (OR), () (grouping), and +ref (inline another set). Definition order is used for tie-breaking.",
+                    "minProperties": 1,
+                    "additionalProperties": {
+                        "type": "string"
+                    }
+                }
+            },
+            "additionalProperties": false
         }
     },
     "properties": {
@@ -306,81 +378,68 @@
                     },
                     "timeouts": {
                         "$ref": "#/definitions/timeouts"
+                    },
+                    "capabilities": {
+                        "type": "object",
+                        "properties": {
+                            "in": {
+                                "type": "array",
+                                "minItems": 1,
+                                "uniqueItems": true,
+                                "default": [],
+                                "items": {
+                                    "type": "string",
+                                    "enum": [
+                                        "text",
+                                        "audio",
+                                        "image"
+                                    ]
+                                },
+                                "description": "List of input modalities understood by the model."
+                            },
+                            "out": {
+                                "type": "array",
+                                "minItems": 1,
+                                "uniqueItems": true,
+                                "default": [],
+                                "items": {
+                                    "type": "string",
+                                    "enum": [
+                                        "text",
+                                        "audio",
+                                        "image"
+                                    ]
+                                },
+                                "description": "List of output modalities generated by the model."
+                            },
+                            "tools": {
+                                "type": "boolean",
+                                "default": false,
+                                "description": "Whether the model supports function calling."
+                            },
+                            "reranker": {
+                                "type": "boolean",
+                                "default": false,
+                                "description": "Whether the model supports the /v1/rerank endpoint."
+                            },
+                            "context": {
+                                "type": "integer",
+                                "minimum": 0,
+                                "default": 0,
+                                "description": "Maximum token context length supported by the model."
+                            }
+                        },
+                        "additionalProperties": false,
+                        "description": "Defines what the model accepts for input, output and other metadata. Used in v1/models to inform clients what the model can do. An empty capabilities block (all zero values) is treated as not configured."
                     }
                 }
             }
         },
         "groups": {
-            "type": "object",
-            "additionalProperties": {
-                "type": "object",
-                "required": [
-                    "members"
-                ],
-                "properties": {
-                    "swap": {
-                        "type": "boolean",
-                        "default": true,
-                        "description": "Controls model swapping behaviour within the group. True: only one model runs at a time. False: all models can run together."
-                    },
-                    "exclusive": {
-                        "type": "boolean",
-                        "default": true,
-                        "description": "Controls how the group affects other groups. True: causes all other groups to unload when this group runs a model. False: does not affect other groups."
-                    },
-                    "persistent": {
-                        "type": "boolean",
-                        "default": false,
-                        "description": "Prevents other groups from unloading the models in this group. Does not affect individual model behaviour."
-                    },
-                    "members": {
-                        "type": "array",
-                        "items": {
-                            "type": "string"
-                        },
-                        "description": "Array of model IDs that are members of this group. Model IDs must be defined in models."
-                    }
-                }
-            },
-            "description": "A dictionary of group settings. Provides advanced controls over model swapping behaviour. Model IDs must be defined in models. A model can only be a member of one group. Behaviour controlled via swap, exclusive, persistent."
+            "$ref": "#/definitions/groupsConfig"
         },
         "matrix": {
-            "type": "object",
-            "description": "Solver-based alternative to groups. Declares valid combinations of concurrent models. The solver minimizes eviction cost when swapping. A config must use either groups or matrix, not both.",
-            "required": [
-                "vars",
-                "sets"
-            ],
-            "properties": {
-                "vars": {
-                    "type": "object",
-                    "description": "Short names for models. Keys must be alphanumeric, 1-8 characters. All sets and evict_costs must use these IDs.",
-                    "minProperties": 1,
-                    "additionalProperties": {
-                        "type": "string"
-                    },
-                    "propertyNames": {
-                        "pattern": "^[a-zA-Z0-9]{1,8}$"
-                    }
-                },
-                "evict_costs": {
-                    "type": "object",
-                    "description": "Relative cost of evicting a running model. Models not listed default to 1. Values must be positive integers.",
-                    "additionalProperties": {
-                        "type": "integer",
-                        "minimum": 1
-                    }
-                },
-                "sets": {
-                    "type": "object",
-                    "description": "Named sets of concurrent model combinations. Values are DSL strings using & (AND), | (OR), () (grouping), and +ref (inline another set). Definition order is used for tie-breaking.",
-                    "minProperties": 1,
-                    "additionalProperties": {
-                        "type": "string"
-                    }
-                }
-            },
-            "additionalProperties": false
+            "$ref": "#/definitions/matrixConfig"
         },
         "hooks": {
             "type": "object",
@@ -512,28 +571,124 @@
             },
             "default": {},
             "description": "A dictionary of remote peers and models they provide. Peers can be another llama-swap or any server that provides the /v1/ generative API endpoints supported by llama-swap."
+        },
+        "upstream": {
+            "type": "object",
+            "description": "Controls behaviour of the /upstream passthrough endpoint. Recommended to only use in special use cases; leaving it as the default will typically be the best experience.",
+            "properties": {
+                "ignorePaths": {
+                    "type": "array",
+                    "items": {
+                        "type": "string"
+                    },
+                    "default": [
+                        ".*\\.(js|json|css|png|gif|jpg|jpeg|ico|txt)$"
+                    ],
+                    "description": "List of RE2 compatible regular expressions. Any request to a path matching any of the regular expressions will be ignored and not trigger a swap. When not specified, defaults to a pattern matching common static-asset suffixes (.js, .json, .css, .png, .gif, .jpg, .jpeg, .ico, .txt)."
+                }
+            },
+            "additionalProperties": false,
+            "default": {}
+        },
+        "routing": {
+            "type": "object",
+            "description": "Canonical routing/scheduling configuration. Alternative to the legacy top-level 'groups'/'matrix' keys; a config must not use both styles.",
+            "properties": {
+                "scheduler": {
+                    "type": "object",
+                    "description": "Scheduler configuration. Decides the order in which queued requests are serviced.",
+                    "properties": {
+                        "use": {
+                            "type": "string",
+                            "enum": [
+                                "serial",
+                                "fifo"
+                            ],
+                            "default": "serial",
+                            "description": "Scheduler to use. 'serial' (default on this fork): strict one-model-at-a-time, requests run in exact arrival order, switching models evicts every other model first. 'fifo': throughput-oriented, batches same-model requests and allows parallel/co-resident models."
+                        },
+                        "settings": {
+                            "type": "object",
+                            "properties": {
+                                "fifo": {
+                                    "type": "object",
+                                    "properties": {
+                                        "priority": {
+                                            "type": "object",
+                                            "description": "Per-model priority. Keys are model IDs, values are integers (default 0). Higher values are serviced first.",
+                                            "additionalProperties": {
+                                                "type": "integer"
+                                            }
+                                        }
+                                    },
+                                    "additionalProperties": false
+                                }
+                            },
+                            "additionalProperties": false
+                        }
+                    },
+                    "additionalProperties": false
+                },
+                "router": {
+                    "type": "object",
+                    "description": "Router configuration. Selects between the group and matrix swapping strategies.",
+                    "properties": {
+                        "use": {
+                            "type": "string",
+                            "enum": [
+                                "group",
+                                "matrix"
+                            ],
+                            "default": "group",
+                            "description": "Router to use. 'group' uses static groups, 'matrix' uses the solver-based swap matrix."
+                        },
+                        "settings": {
+                            "type": "object",
+                            "properties": {
+                                "groups": {
+                                    "$ref": "#/definitions/groupsConfig"
+                                },
+                                "matrix": {
+                                    "$ref": "#/definitions/matrixConfig"
+                                }
+                            },
+                            "additionalProperties": false
+                        }
+                    },
+                    "additionalProperties": false
+                }
+            },
+            "additionalProperties": false
         }
     },
     "allOf": [
         {
             "if": {
-                "required": ["groups"]
+                "required": [
+                    "groups"
+                ]
             },
             "then": {
                 "not": {
-                    "required": ["matrix"]
+                    "required": [
+                        "matrix"
+                    ]
                 }
             }
         },
         {
             "if": {
-                "required": ["matrix"]
+                "required": [
+                    "matrix"
+                ]
             },
             "then": {
                 "not": {
-                    "required": ["groups"]
+                    "required": [
+                        "groups"
+                    ]
                 }
             }
         }
     ]
-}
+}

config.example.yaml

@@ -134,6 +134,18 @@ apiKeys:
   - "${env.API_KEY_1}"
   - "${env.API_KEY_2}"
 
+# upstream: controls behaviour of the /upstream passthrough endpoint
+# - optional, default: empty dictionary
+# - recommended to only use in special use cases. Leaving it as the
+#   default will typically be the best experience
+upstream:
+  # ignorePaths: list of RE2 compatible regular expressions
+  # - default: (see below)
+  # - any request to a path matching any of the regular expressions
+  #   will be ignored and not trigger a swap
+  ignorePaths:
+    - '.*\.(js|json|css|png|gif|jpg|jpeg|ico|txt)$'
+
 # models: a dictionary of model configurations
 # - required
 # - each key is the model's ID, used in API requests
@@ -312,6 +324,37 @@ models:
       tlsHandshake: 10
       idleConn: 90
 
+    # capabilities: defines what the model accepts for input, output and other metadata
+    # - optional; omitted or all-zero means no capabilities
+    # - used in v1/models to inform clients what the model can do
+    capabilities:
+      # in: list of modalities understood by the model
+      # - default: []
+      # - valid: text, audio, image
+      in:
+        - text
+        - audio
+        - image
+      # out: list of modalities generated by the model
+      # - default: []
+      # - valid: text, audio, image
+      out:
+        - text
+        - audio
+        - image
+      # tools: the model supports function calling
+      # - default: false
+      tools: true
+
+      # reranker: the model supports the /v1/rerank endpoint
+      # - default: false
+      reranker: false
+
+      # context: the maximum token context length supported
+      # - default: 0
+      # - must be an integer > 0
+      context: 32000
+
   # Unlisted model example:
   "qwen-unlisted":
     # unlisted: boolean, true or false
@@ -343,93 +386,6 @@ models:
     # - processes have 5 seconds to shutdown until forceful termination is attempted
     cmdStop: docker stop ${MODEL_ID}
 
-# =============================================================================
-# matrix: run concurrent models with a solver-based swap DSL
-# =============================================================================
-#
-# Matrix or Groups?
-#
-# Groups are available and fully supported. The syntax may be easier to use
-# for simple use cases.
-#
-# Documentation can be found here:
-# https://github.com/mostlygeek/llama-swap/blob/40e39f7/config.example.yaml#L334-L396
-#
-# A config can only use a matrix (recommended) or groups. A configuration error
-# will occur if both are defined. Groups is legacy but is fully supported with
-# no plans to deprecate it.
-#
-# ~~~~~
-#
-# The matrix declares valid combinations of models that can run concurrently.
-# When a model is requested, the solver finds the cheapest way to make it
-# available by evicting as few (and least costly) running models as possible.
-#
-# Solver behavior:
-#   1. Request arrives for model X
-#   2. If X is already running, forward immediately. Done.
-#   3. Find all sets containing X
-#   4. For each candidate set, compute cost: sum of evict_costs for
-#      every running model NOT in that set
-#   5. Pick lowest cost candidate. Ties broken by definition order.
-#   6. Evict what needs to stop. Start X. Forward request.
-#
-# Subset semantics: a set [a, b, c] means any subset is valid.
-# Only the requested model is started — others are not preloaded.
-#
-# A model not appearing in any set can only run alone.
-#
-matrix:
-  # vars: short names for models (alphanumeric, 1-8 chars)
-  # - required for sets and evict_costs settings
-  # - each entry is a short name to a real model ID. Do not use an alias
-  # - used to keep set DSL logic short and easier to read
-  # - sets and evict_costs only use identifiers defined in vars
-  vars:
-    g: gemma-model
-    q: qwen-model
-    m: mistral-model
-    v: voxtral-model
-    e: reranker-model
-    L: llama-70B
-    sd: stable-diffusion
-
-  # evict_costs: relative cost of losing a running model (default: 1)
-  evict_costs:
-    v: 50 # vllm backend, slow cold start
-    L: 30 # 70B weights, slow to load
-
-  # sets: named sets of concurrent model combinations
-  # Values are DSL strings with operators:
-  #   &     AND (models run together)
-  #   |     OR  (alternatives)
-  #   ()    grouping
-  #   +ref  inline another set's expression
-  #
-  # Expansion examples:
-  #   "L"                  → [L]
-  #   "a & b"              → [a, b]
-  #   "a | b"              → [a], [b]
-  #   "(a | b) & c"        → [a, c], [b, c]
-  #   "(a | b) & (c | d)"  → [a,c], [a,d], [b,c], [b,d]
-  #   "+llms & v"          → expands llms inline, then applies & v
-  sets:
-    # LLM + TTS: switching between g/q/m won't evict v
-    # expands to: [g,v], [q,v], [m,v]
-    standard: "(g | q | m) & v"
-
-    # LLM + TTS + reranker
-    # expands to: [g,v,e], [q,v,e]
-    with_rerank: "(g | q) & v & e"
-
-    # LLM + image generation, no TTS
-    # expands to: [g,sd], [q,sd]
-    creative: "(g | q) & sd"
-
-    # 70B model uses all GPUs, can only run alone
-    # expands to: [L]
-    full: "L"
-
 # hooks: a dictionary of event triggers and actions
 # - optional, default: empty dictionary
 # - the only supported hook is on_startup
@@ -446,6 +402,186 @@ hooks:
     preload:
       - "llama"
 
+# routing:
+# Controls how llama-swap decides which models can run at the same time and
+# which get swapped out. Choose one of two swap engines:
+#
+# - group:  the default engine. Simpler to configure. You define groups of
+#           models that run together, and loading one group typically unloads
+#           the others.
+#
+# - matrix: the newer engine. More involved to configure, but far more
+#           flexible. It uses a small expression language to describe which
+#           model combinations are allowed to run concurrently, enabling
+#           setups that groups cannot express.
+#
+# The routing section is optional.
+routing:
+  router:
+    # use: a string defining which engine to use
+    # - optional, default: "group"
+    # - valid values: group, matrix
+    use: group
+
+    # settings: a dictionary of settings for the specific engines
+    settings:
+      # groups: a dictionary of named groups
+      # - optional, default: empty dictionary
+      # - lets you keep some models loaded while others swap out
+      # - every member must be a model ID defined in the models section
+      # - a model can belong to only one group
+      # - behaviour is set per group with the `swap`, `exclusive` and
+      #   `persistent` fields
+      # - see issue #109 for details
+      #
+      # NOTE: the model names below are illustrative and are not defined above.
+      groups:
+        # group1 reproduces llama-swap's default behaviour: only one model
+        # runs at a time across the entire instance.
+        "group1":
+          # swap: how members of this group swap among themselves
+          # - optional, default: true
+          # - true:  only one member runs at a time
+          # - false: all members can run together, no swapping
+          swap: true
+
+          # exclusive: how this group affects other groups
+          # - optional, default: true
+          # - true:  running a member unloads every other group
+          # - false: running a member leaves other groups untouched
+          exclusive: true
+
+          # members: the model IDs in this group
+          # required
+          members:
+            - "llama"
+            - "qwen-unlisted"
+
+        # group2: members all run together, but loading any other group
+        # unloads them.
+        "group2":
+          # swap: false lets all members stay loaded at once
+          swap: false
+
+          # exclusive: false means requesting a member loads it without
+          # unloading any other group
+          exclusive: false
+          members:
+            - "docker-llama"
+            - "modelA"
+            - "modelB"
+
+        # forever: a persistent group that other groups can never unload.
+        "forever":
+          # persistent: other groups cannot unload this group's members
+          # - optional, default: false
+          # - has no effect on swapping within the group
+          persistent: true
+
+          # swap/exclusive: false keeps all members loaded and avoids
+          # unloading other groups
+          swap: false
+          exclusive: false
+          members:
+            - "forever-modelA"
+            - "forever-modelB"
+            - "forever-modelc"
+
+      # The matrix lists the model combinations that are allowed to run
+      # concurrently. When a model is requested, the solver makes room for it
+      # by evicting as few running models as possible, preferring to keep the
+      # costliest ones loaded.
+      #
+      # Solver behaviour:
+      #   1. A request arrives for model X.
+      #   2. If X is already running, forward the request. Done.
+      #   3. Collect every set that contains X.
+      #   4. For each set, add up the evict_costs of the running models that
+      #      are NOT in that set — that is the set's cost.
+      #   5. Choose the lowest-cost set. Break ties by definition order.
+      #   6. Evict the models outside that set, start X, forward the request.
+      #
+      # Subset semantics: a set [a, b, c] also permits any subset of itself.
+      # Only the requested model is started; the others are not preloaded.
+      #
+      # A model that appears in no set can only run on its own.
+      #
+      matrix:
+        # vars: short aliases for model IDs (alphanumeric, 1-8 chars)
+        # - required: sets and evict_costs reference these names, not model IDs
+        # - map each short name to a real model ID (not a model alias)
+        # - keeps the set expressions short and readable
+        vars:
+          g: gemma-model
+          q: qwen-model
+          m: mistral-model
+          v: voxtral-model
+          e: reranker-model
+          L: llama-70B
+          sd: stable-diffusion
+
+        # evict_costs: relative cost of losing a running model (default: 1)
+        evict_costs:
+          v: 50 # vllm backend, slow cold start
+          L: 30 # 70B weights, slow to load
+
+        # sets: named combinations of models that may run together.
+        # Each value is an expression built from these operators:
+        #   &     AND   (models run together)
+        #   |     OR    (alternatives)
+        #   ()    grouping
+        #   +ref  inline the expression of another set
+        #
+        # Each expression expands into one or more concrete sets:
+        #   "L"                  → [L]
+        #   "a & b"              → [a, b]
+        #   "a | b"              → [a], [b]
+        #   "(a | b) & c"        → [a, c], [b, c]
+        #   "(a | b) & (c | d)"  → [a,c], [a,d], [b,c], [b,d]
+        #   "+llms & v"          → inline the llms set, then AND with v
+        sets:
+          # An LLM plus TTS. Switching between g/q/m keeps v loaded.
+          # expands to: [g,v], [q,v], [m,v]
+          standard: "(g | q | m) & v"
+
+          # An LLM plus TTS plus reranker.
+          # expands to: [g,v,e], [q,v,e]
+          with_rerank: "(g | q) & v & e"
+
+          # An LLM plus image generation, no TTS.
+          # expands to: [g,sd], [q,sd]
+          creative: "(g | q) & sd"
+
+          # The 70B model uses every GPU, so it can only run alone.
+          # expands to: [L]
+          full: "L"
+
+  # scheduler: how queued requests are ordered and run.
+  # - optional, default on this fork: "serial"
+  # - valid values:
+  #   - "serial": strict one-model-at-a-time. Requests run in exact arrival
+  #     order; only one request runs at a time; switching to a different model
+  #     evicts every other running model first so a single model occupies memory
+  #     at a time. This ignores group/matrix co-residency entirely. The "fifo"
+  #     settings below (priority) do not apply.
+  #   - "fifo": throughput-oriented. Same-model requests are batched to reduce
+  #     swaps and a model serves up to its concurrencyLimit in parallel; models
+  #     in non-exclusive groups can run concurrently. Requests may be reordered.
+  scheduler:
+    use: serial
+    settings:
+      # fifo settings only apply when use: fifo
+      fifo:
+        # priority: a dictionary of model ID -> priority
+        # - optional, default: empty dictionary
+        # - models default to priority 0
+        # - higher priority requests are serviced first in the queue
+        priority:
+          A: 10
+          B: 5
+          C: 5
+          D: 1
+
 # peers: a dictionary of remote peers and models they provide
 # - optional, default empty dictionary
 # - peers can be another llama-swap

docker/fork-cuda.Containerfile

@@ -0,0 +1,74 @@
+# Build a CUDA llama-swap image FROM THIS FORK's source (includes the serial
+# scheduler) and layer it on a pinned llama.cpp CUDA server base. Produces e.g.:
+#   gitea.stevedudenhoeffer.com/steve/llama-swap:v230-cuda-b9821
+#
+# BASE_TAG selects the llama.cpp CUDA runtime + llama-server build, e.g.
+# "server-cuda-b9821". The llama-swap binary (with the embedded Svelte UI) is
+# compiled from the repo at build time, so no GitHub release is required.
+#
+# Build context is the repo root:
+#   docker build -f docker/fork-cuda.Containerfile \
+#     --build-arg BASE_TAG=server-cuda-b9821 -t llama-swap:v230-cuda-b9821 .
+
+ARG BASE_IMAGE=ghcr.io/ggml-org/llama.cpp
+ARG BASE_TAG=server-cuda-b9821
+
+# ---- Stage 1: build the Svelte UI (embedded into the binary) ----
+FROM node:22-bookworm-slim AS ui
+WORKDIR /src/ui-svelte
+# Install deps first for layer caching. .npmrc carries legacy-peer-deps=true,
+# which the project relies on (tailwind/vite peer ranges), so copy it before
+# npm ci or the strict resolver fails with ERESOLVE.
+COPY ui-svelte/package.json ui-svelte/package-lock.json ui-svelte/.npmrc ./
+RUN npm ci
+COPY ui-svelte/ ./
+# `npm run build` is `vite build --emptyOutDir`; vite.config.ts writes to
+# ../internal/server/ui_dist, which //go:embed picks up in the next stage.
+RUN mkdir -p /src/internal/server && npm run build
+
+# ---- Stage 2: build the llama-swap binary with the embedded UI ----
+FROM golang:1.26-bookworm AS build
+WORKDIR /src
+# Cache modules independently of source churn.
+COPY go.mod go.sum ./
+RUN go mod download
+COPY . .
+# Overlay the freshly built UI so //go:embed ui_dist ships the real assets
+# instead of the committed placeholder.
+COPY --from=ui /src/internal/server/ui_dist/ ./internal/server/ui_dist/
+ARG LS_VERSION=v230
+ARG GIT_HASH=unknown
+ARG BUILD_DATE=unknown
+RUN CGO_ENABLED=0 GOOS=linux go build \
+    -ldflags="-X main.version=${LS_VERSION} -X main.commit=${GIT_HASH} -X main.date=${BUILD_DATE}" \
+    -o /out/llama-swap .
+
+# ---- Stage 3: runtime image on the pinned llama.cpp CUDA base ----
+FROM ${BASE_IMAGE}:${BASE_TAG}
+
+# Run as root by default to match the upstream `vNNN-cuda-bNNNN` (non-suffixed)
+# image that ragnaros pulls today: it needs root to reach the mounted docker
+# socket for container-backed models (sd-server). Override UID/GID at build time
+# for a non-root variant.
+ARG UID=0
+ARG GID=0
+ARG USER_HOME=/root
+ENV HOME=$USER_HOME
+
+RUN set -eux; \
+    if [ "$UID" -ne 0 ]; then \
+      if [ "$GID" -ne 0 ]; then groupadd --system --gid "$GID" app; fi; \
+      useradd --system --uid "$UID" --gid "$GID" --home "$USER_HOME" app; \
+    fi; \
+    mkdir --parents "$HOME" /app; \
+    chown --recursive "$UID:$GID" "$HOME" /app
+
+COPY --from=build --chown=$UID:$GID /out/llama-swap /app/llama-swap
+COPY --chown=$UID:$GID docker/config.example.yaml /app/config.yaml
+
+USER $UID:$GID
+WORKDIR /app
+ENV PATH="/app:${PATH}"
+
+HEALTHCHECK CMD curl -f http://localhost:8080/ || exit 1
+ENTRYPOINT [ "/app/llama-swap", "-config", "/app/config.yaml" ]

docker/llama-swap.Containerfile

@@ -2,10 +2,6 @@ ARG BASE_IMAGE=ghcr.io/ggml-org/llama.cpp
 ARG BASE_TAG=server-cuda
 FROM ${BASE_IMAGE}:${BASE_TAG}
 
-# has to be after the FROM
-# TARGETARCH is auto-set by `docker buildx build --platform …` (amd64/arm64);
-# falls back to amd64 when an older `docker build` runs without buildx.
-ARG TARGETARCH=amd64
 ARG LS_VER=170
 ARG LS_REPO=mostlygeek/llama-swap
 
@@ -37,9 +33,15 @@ WORKDIR /app
 ENV PATH="/app:${PATH}"
 
 RUN \
-    curl -LO "https://github.com/${LS_REPO}/releases/download/v${LS_VER}/llama-swap_${LS_VER}_linux_${TARGETARCH}.tar.gz" && \
-    tar -zxf "llama-swap_${LS_VER}_linux_${TARGETARCH}.tar.gz" && \
-    rm "llama-swap_${LS_VER}_linux_${TARGETARCH}.tar.gz"
+    set -eux; \
+    case "$(uname -m)" in \
+        x86_64)  ARCH=amd64 ;; \
+        aarch64) ARCH=arm64 ;; \
+        *) echo "unsupported arch: $(uname -m)" >&2; exit 1 ;; \
+    esac; \
+    curl --fail -LO "https://github.com/${LS_REPO}/releases/download/v${LS_VER}/llama-swap_${LS_VER}_linux_${ARCH}.tar.gz" && \
+    tar -zxf "llama-swap_${LS_VER}_linux_${ARCH}.tar.gz" && \
+    rm "llama-swap_${LS_VER}_linux_${ARCH}.tar.gz"
 
 COPY --chown=$UID:$GID config.example.yaml /app/config.yaml
 

go.mod

@@ -9,12 +9,14 @@ require (
 	github.com/charmbracelet/lipgloss v1.1.0
 	github.com/fxamacker/cbor/v2 v2.9.1
 	github.com/gin-gonic/gin v1.10.0
+	github.com/google/jsonschema-go v0.4.3
 	github.com/klauspost/compress v1.18.5
 	github.com/shirou/gopsutil/v4 v4.26.4
 	github.com/stretchr/testify v1.11.1
 	github.com/tidwall/gjson v1.18.0
 	github.com/tidwall/sjson v1.2.5
 	golang.org/x/sync v0.20.0
+	golang.org/x/sys v0.41.0
 	gopkg.in/yaml.v3 v3.0.1
 )
 
@@ -70,7 +72,6 @@ require (
 	golang.org/x/arch v0.8.0 // indirect
 	golang.org/x/crypto v0.45.0 // indirect
 	golang.org/x/net v0.47.0 // indirect
-	golang.org/x/sys v0.41.0 // indirect
 	golang.org/x/text v0.31.0 // indirect
 	google.golang.org/protobuf v1.34.1 // indirect
 )

go.sum

@@ -61,6 +61,8 @@ github.com/google/go-cmp v0.5.6/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/
 github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
 github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
+github.com/google/jsonschema-go v0.4.3 h1:/DBOLZTfDow7pe2GmaJNhltueGTtDKICi8V8p+DQPd0=
+github.com/google/jsonschema-go v0.4.3/go.mod h1:r5quNTdLOYEz95Ru18zA0ydNbBuYoo9tgaYcxEYhJVE=
 github.com/json-iterator/go v1.1.12 h1:PV8peI4a0ysnczrg+LtxykD8LfKY9ML6u2jnxaEnrnM=
 github.com/json-iterator/go v1.1.12/go.mod h1:e30LSqwooZae/UwlEbR2852Gd8hjQvJoHmT4TnhNGBo=
 github.com/klauspost/compress v1.18.5 h1:/h1gH5Ce+VWNLSWqPzOVn6XBO+vJbCNGvjoaGBFW2IE=

internal/config/commands.go

@@ -0,0 +1,57 @@
+package config
+
+import (
+	"fmt"
+	"runtime"
+	"strings"
+
+	"github.com/billziss-gh/golib/shlex"
+)
+
+func SanitizeCommand(cmdStr string) ([]string, error) {
+	var cleanedLines []string
+	for _, line := range strings.Split(cmdStr, "\n") {
+		trimmed := strings.TrimSpace(line)
+		// Skip comment lines
+		if strings.HasPrefix(trimmed, "#") {
+			continue
+		}
+		// Handle trailing backslashes by replacing with space
+		if strings.HasSuffix(trimmed, "\\") {
+			cleanedLines = append(cleanedLines, strings.TrimSuffix(trimmed, "\\")+" ")
+		} else {
+			cleanedLines = append(cleanedLines, line)
+		}
+	}
+
+	// put it back together
+	cmdStr = strings.Join(cleanedLines, "\n")
+
+	// Split the command into arguments
+	var args []string
+	if runtime.GOOS == "windows" {
+		args = shlex.Windows.Split(cmdStr)
+	} else {
+		args = shlex.Posix.Split(cmdStr)
+	}
+
+	// Ensure the command is not empty
+	if len(args) == 0 {
+		return nil, fmt.Errorf("empty command")
+	}
+
+	return args, nil
+}
+
+func StripComments(cmdStr string) string {
+	var cleanedLines []string
+	for _, line := range strings.Split(cmdStr, "\n") {
+		trimmed := strings.TrimSpace(line)
+		// Skip comment lines
+		if strings.HasPrefix(trimmed, "#") {
+			continue
+		}
+		cleanedLines = append(cleanedLines, line)
+	}
+	return strings.Join(cleanedLines, "\n")
+}

internal/config/config.go

@@ -2,16 +2,9 @@ package config
 
 import (
 	"fmt"
-	"io"
-	"net/url"
 	"os"
-	"regexp"
-	"runtime"
 	"sort"
-	"strings"
-	"time"
 
-	"github.com/billziss-gh/golib/shlex"
 	"gopkg.in/yaml.v3"
 )
 
@@ -85,12 +78,6 @@ type GroupConfig struct {
 	Members    []string `yaml:"members"`
 }
 
-var (
-	macroNameRegex    = regexp.MustCompile(`^[a-zA-Z0-9_-]+$`)
-	macroPatternRegex = regexp.MustCompile(`\$\{([a-zA-Z0-9_-]+)\}`)
-	envMacroRegex     = regexp.MustCompile(`\$\{env\.([a-zA-Z_][a-zA-Z0-9_]*)\}`)
-)
-
 // set default values for GroupConfig
 func (c *GroupConfig) UnmarshalYAML(unmarshal func(interface{}) error) error {
 	type rawGroupConfig GroupConfig
@@ -129,13 +116,16 @@ type Config struct {
 	GlobalTTL          int                    `yaml:"globalTTL"`
 	Models             map[string]ModelConfig `yaml:"models"` /* key is model ID */
 	Profiles           map[string][]string    `yaml:"profiles"`
-	Groups             map[string]GroupConfig `yaml:"groups"` /* key is group ID */
 
-	// swap matrix: solver-based alternative to groups
-	Matrix *MatrixConfig `yaml:"matrix"`
+	// routing is the canonical source for swap/scheduling configuration.
+	// New code must read Routing, never the backwards-compat fields below.
+	Routing RoutingConfig `yaml:"routing"`
 
-	// populated during validation when matrix is configured
-	ExpandedSets []ExpandedSet `yaml:"-"`
+	// Groups and Matrix are permanent backwards-compat input fields for the
+	// legacy top-level `groups:`/`matrix:` keys. They are normalized into
+	// Routing by LoadConfigFromReader. New code must not read them directly.
+	Groups map[string]GroupConfig `yaml:"groups"` /* key is group ID */
+	Matrix *MatrixConfig          `yaml:"matrix"`
 
 	// for key/value replacements in model's cmd, cmdStop, proxy, checkEndPoint
 	Macros MacroList `yaml:"macros"`
@@ -160,6 +150,38 @@ type Config struct {
 
 	// support remote peers, see issue #433, #296
 	Peers PeerDictionaryConfig `yaml:"peers"`
+
+	// upstream controls behaviour of the /upstream passthrough endpoint
+	Upstream UpstreamConfig `yaml:"upstream"`
+}
+
+// RoutingConfig is the canonical, normalized routing/scheduling configuration.
+type RoutingConfig struct {
+	Scheduler SchedulerConfig `yaml:"scheduler"`
+	Router    RouterConfig    `yaml:"router"`
+}
+
+type SchedulerConfig struct {
+	Use      string            `yaml:"use"` // default "fifo"
+	Settings SchedulerSettings `yaml:"settings"`
+}
+
+type SchedulerSettings struct {
+	Fifo FifoConfig `yaml:"fifo"`
+}
+
+type FifoConfig struct {
+	Priority map[string]int `yaml:"priority"` // model ID -> priority, default 0
+}
+
+type RouterConfig struct {
+	Use      string         `yaml:"use"` // "group" (default) | "matrix"
+	Settings RouterSettings `yaml:"settings"`
+}
+
+type RouterSettings struct {
+	Groups map[string]GroupConfig `yaml:"groups"`
+	Matrix *MatrixConfig          `yaml:"matrix"`
 }
 
 func (c *Config) RealModelName(search string) (string, bool) {
@@ -189,369 +211,6 @@ func LoadConfig(path string) (Config, error) {
 	return LoadConfigFromReader(file)
 }
 
-func LoadConfigFromReader(r io.Reader) (Config, error) {
-	data, err := io.ReadAll(r)
-	if err != nil {
-		return Config{}, err
-	}
-	yamlStr := string(data)
-
-	// Phase 1: Substitute all ${env.VAR} macros at string level
-	// This is safe because env values are simple strings without YAML formatting
-	yamlStr, err = substituteEnvMacros(yamlStr)
-	if err != nil {
-		return Config{}, err
-	}
-
-	// Unmarshal into full Config with defaults
-	config := Config{
-		HealthCheckTimeout: 120,
-		StartPort:          5800,
-		LogLevel:           "info",
-		LogTimeFormat:      "",
-		LogToStdout:        LogToStdoutProxy,
-		MetricsMaxInMemory: 1000,
-		CaptureBuffer:      5,
-		GlobalTTL:          0,
-	}
-	if err = yaml.Unmarshal([]byte(yamlStr), &config); err != nil {
-		return Config{}, err
-	}
-
-	if config.HealthCheckTimeout < 15 {
-		config.HealthCheckTimeout = 15
-	}
-
-	// Apply defaults for performance config when section is missing
-	if config.Performance.Every == 0 {
-		config.Performance.Every = 5 * time.Second
-	}
-	if err = config.Performance.Validate(); err != nil {
-		return Config{}, fmt.Errorf("performance: %w", err)
-	}
-
-	if config.StartPort < 1 {
-		return Config{}, fmt.Errorf("startPort must be greater than 1")
-	}
-
-	if config.GlobalTTL < 0 {
-		return Config{}, fmt.Errorf("globalTTL must be >= 0")
-	}
-
-	switch config.LogToStdout {
-	case LogToStdoutProxy, LogToStdoutUpstream, LogToStdoutBoth, LogToStdoutNone:
-	default:
-		return Config{}, fmt.Errorf("logToStdout must be one of: proxy, upstream, both, none")
-	}
-
-	// Populate the aliases map
-	config.aliases = make(map[string]string)
-	for modelName, modelConfig := range config.Models {
-		for _, alias := range modelConfig.Aliases {
-			if _, found := config.aliases[alias]; found {
-				return Config{}, fmt.Errorf("duplicate alias %s found in model: %s", alias, modelName)
-			}
-			config.aliases[alias] = modelName
-		}
-	}
-
-	// Validate global macros
-	for _, macro := range config.Macros {
-		if err = validateMacro(macro.Name, macro.Value); err != nil {
-			return Config{}, err
-		}
-	}
-
-	// Get and sort all model IDs for consistent port assignment
-	modelIds := make([]string, 0, len(config.Models))
-	for modelId := range config.Models {
-		modelIds = append(modelIds, modelId)
-	}
-	sort.Strings(modelIds)
-
-	nextPort := config.StartPort
-	for _, modelId := range modelIds {
-		modelConfig := config.Models[modelId]
-		modelConfig.HealthCheckTimeout = config.HealthCheckTimeout
-
-		// Strip comments from command fields
-		modelConfig.Cmd = StripComments(modelConfig.Cmd)
-		modelConfig.CmdStop = StripComments(modelConfig.CmdStop)
-
-		// set model TTL to globalTTL it is the default value
-		if modelConfig.UnloadAfter == MODEL_CONFIG_DEFAULT_TTL {
-			modelConfig.UnloadAfter = config.GlobalTTL
-		}
-
-		if modelConfig.UnloadAfter < 0 {
-			return Config{}, fmt.Errorf("model %s: invalid TTL value %d", modelId, modelConfig.UnloadAfter)
-		}
-
-		// Validate model macros
-		for _, macro := range modelConfig.Macros {
-			if err = validateMacro(macro.Name, macro.Value); err != nil {
-				return Config{}, fmt.Errorf("model %s: %s", modelId, err.Error())
-			}
-		}
-
-		// Build merged macro list: MODEL_ID + global macros + model macros (model overrides global)
-		mergedMacros := make(MacroList, 0, len(config.Macros)+len(modelConfig.Macros)+1)
-		mergedMacros = append(mergedMacros, MacroEntry{Name: "MODEL_ID", Value: modelId})
-		mergedMacros = append(mergedMacros, config.Macros...)
-
-		// Add model macros (override globals with same name)
-		for _, entry := range modelConfig.Macros {
-			found := false
-			for i, existing := range mergedMacros {
-				if existing.Name == entry.Name {
-					mergedMacros[i] = entry
-					found = true
-					break
-				}
-			}
-			if !found {
-				mergedMacros = append(mergedMacros, entry)
-			}
-		}
-
-		// Substitute remaining macros in model fields (LIFO order)
-		for i := len(mergedMacros) - 1; i >= 0; i-- {
-			entry := mergedMacros[i]
-			macroSlug := fmt.Sprintf("${%s}", entry.Name)
-			macroStr := fmt.Sprintf("%v", entry.Value)
-
-			modelConfig.Cmd = strings.ReplaceAll(modelConfig.Cmd, macroSlug, macroStr)
-			modelConfig.CmdStop = strings.ReplaceAll(modelConfig.CmdStop, macroSlug, macroStr)
-			modelConfig.Proxy = strings.ReplaceAll(modelConfig.Proxy, macroSlug, macroStr)
-			modelConfig.CheckEndpoint = strings.ReplaceAll(modelConfig.CheckEndpoint, macroSlug, macroStr)
-			modelConfig.Filters.StripParams = strings.ReplaceAll(modelConfig.Filters.StripParams, macroSlug, macroStr)
-			modelConfig.Name = strings.ReplaceAll(modelConfig.Name, macroSlug, macroStr)
-			modelConfig.Description = strings.ReplaceAll(modelConfig.Description, macroSlug, macroStr)
-
-			// Substitute macros in SetParamsByID keys and values
-			if len(modelConfig.Filters.SetParamsByID) > 0 {
-				newSetParamsByID := make(map[string]map[string]any, len(modelConfig.Filters.SetParamsByID))
-				for key, paramMap := range modelConfig.Filters.SetParamsByID {
-					newKey := strings.ReplaceAll(key, macroSlug, macroStr)
-					newValAny, err := substituteMacroInValue(any(paramMap), entry.Name, entry.Value)
-					if err != nil {
-						return Config{}, fmt.Errorf("model %s filters.setParamsByID: %s", modelId, err.Error())
-					}
-					newParamMap, ok := newValAny.(map[string]any)
-					if !ok {
-						return Config{}, fmt.Errorf("model %s filters.setParamsByID: unexpected type after macro substitution", modelId)
-					}
-					newSetParamsByID[newKey] = newParamMap
-				}
-				modelConfig.Filters.SetParamsByID = newSetParamsByID
-			}
-
-			// Substitute in metadata (type-preserving)
-			if len(modelConfig.Metadata) > 0 {
-				result, err := substituteMacroInValue(modelConfig.Metadata, entry.Name, entry.Value)
-				if err != nil {
-					return Config{}, fmt.Errorf("model %s metadata: %s", modelId, err.Error())
-				}
-				modelConfig.Metadata = result.(map[string]any)
-			}
-		}
-
-		// Handle PORT macro - only allocate if cmd uses it
-		cmdHasPort := strings.Contains(modelConfig.Cmd, "${PORT}")
-		proxyHasPort := strings.Contains(modelConfig.Proxy, "${PORT}")
-		if cmdHasPort || proxyHasPort {
-			if !cmdHasPort && proxyHasPort {
-				return Config{}, fmt.Errorf("model %s: proxy uses ${PORT} but cmd does not - ${PORT} is only available when used in cmd", modelId)
-			}
-
-			macroSlug := "${PORT}"
-			macroStr := fmt.Sprintf("%v", nextPort)
-
-			modelConfig.Cmd = strings.ReplaceAll(modelConfig.Cmd, macroSlug, macroStr)
-			modelConfig.CmdStop = strings.ReplaceAll(modelConfig.CmdStop, macroSlug, macroStr)
-			modelConfig.Proxy = strings.ReplaceAll(modelConfig.Proxy, macroSlug, macroStr)
-			modelConfig.Name = strings.ReplaceAll(modelConfig.Name, macroSlug, macroStr)
-			modelConfig.Description = strings.ReplaceAll(modelConfig.Description, macroSlug, macroStr)
-
-			if len(modelConfig.Metadata) > 0 {
-				result, err := substituteMacroInValue(modelConfig.Metadata, "PORT", nextPort)
-				if err != nil {
-					return Config{}, fmt.Errorf("model %s metadata: %s", modelId, err.Error())
-				}
-				modelConfig.Metadata = result.(map[string]any)
-			}
-
-			nextPort++
-		}
-
-		// Validate no unknown macros remain
-		fieldMap := map[string]string{
-			"cmd":                 modelConfig.Cmd,
-			"cmdStop":             modelConfig.CmdStop,
-			"proxy":               modelConfig.Proxy,
-			"checkEndpoint":       modelConfig.CheckEndpoint,
-			"filters.stripParams": modelConfig.Filters.StripParams,
-			"name":                modelConfig.Name,
-			"description":         modelConfig.Description,
-		}
-
-		for fieldName, fieldValue := range fieldMap {
-			matches := macroPatternRegex.FindAllStringSubmatch(fieldValue, -1)
-			for _, match := range matches {
-				macroName := match[1]
-				if macroName == "PID" && fieldName == "cmdStop" {
-					continue // replaced at runtime
-				}
-				if macroName == "PORT" || macroName == "MODEL_ID" {
-					return Config{}, fmt.Errorf("macro '${%s}' should have been substituted in %s.%s", macroName, modelId, fieldName)
-				}
-				return Config{}, fmt.Errorf("unknown macro '${%s}' found in %s.%s", macroName, modelId, fieldName)
-			}
-		}
-
-		if len(modelConfig.Metadata) > 0 {
-			if err := validateNestedForUnknownMacros(modelConfig.Metadata, fmt.Sprintf("model %s metadata", modelId)); err != nil {
-				return Config{}, err
-			}
-		}
-
-		// Validate SetParamsByID keys and values
-		for key, paramMap := range modelConfig.Filters.SetParamsByID {
-			if matches := macroPatternRegex.FindAllStringSubmatch(key, -1); len(matches) > 0 {
-				return Config{}, fmt.Errorf("unknown macro '${%s}' found in model %s filters.setParamsByID key", matches[0][1], modelId)
-			}
-			if err := validateNestedForUnknownMacros(any(paramMap), fmt.Sprintf("model %s filters.setParamsByID[%s]", modelId, key)); err != nil {
-				return Config{}, err
-			}
-		}
-
-		// Auto-register setParamsByID keys as aliases (skip the model's own ID)
-		for key := range modelConfig.Filters.SetParamsByID {
-			if key == modelId {
-				continue
-			}
-			if _, exists := config.Models[key]; exists {
-				return Config{}, fmt.Errorf("model %s filters.setParamsByID: key '%s' conflicts with an existing model ID", modelId, key)
-			}
-			if existingModel, exists := config.aliases[key]; exists {
-				if existingModel != modelId {
-					return Config{}, fmt.Errorf("duplicate alias '%s' in model %s filters.setParamsByID, already used by model %s", key, modelId, existingModel)
-				}
-				continue // already registered as explicit alias for this model
-			}
-			config.aliases[key] = modelId
-			modelConfig.Aliases = append(modelConfig.Aliases, key)
-		}
-
-		if _, err := url.Parse(modelConfig.Proxy); err != nil {
-			return Config{}, fmt.Errorf("model %s: invalid proxy URL: %w", modelId, err)
-		}
-
-		if modelConfig.SendLoadingState == nil {
-			v := config.SendLoadingState
-			modelConfig.SendLoadingState = &v
-		}
-
-		config.Models[modelId] = modelConfig
-	}
-
-	// groups XOR matrix
-	if config.Matrix != nil && len(config.Groups) > 0 {
-		return Config{}, fmt.Errorf("config cannot use both 'groups' and 'matrix'")
-	}
-
-	if config.Matrix != nil {
-		expandedSets, err := ValidateMatrix(*config.Matrix, config.Models)
-		if err != nil {
-			return Config{}, fmt.Errorf("matrix: %w", err)
-		}
-		config.ExpandedSets = expandedSets
-	} else {
-		config = AddDefaultGroupToConfig(config)
-
-		// Validate group members
-		memberUsage := make(map[string]string)
-		for groupID, groupConfig := range config.Groups {
-			prevSet := make(map[string]bool)
-			for _, member := range groupConfig.Members {
-				if _, found := prevSet[member]; found {
-					return Config{}, fmt.Errorf("duplicate model member %s found in group: %s", member, groupID)
-				}
-				prevSet[member] = true
-
-				if existingGroup, exists := memberUsage[member]; exists {
-					return Config{}, fmt.Errorf("model member %s is used in multiple groups: %s and %s", member, existingGroup, groupID)
-				}
-				memberUsage[member] = groupID
-			}
-		}
-	}
-
-	// Clean up hooks preload
-	if len(config.Hooks.OnStartup.Preload) > 0 {
-		var toPreload []string
-		for _, modelID := range config.Hooks.OnStartup.Preload {
-			modelID = strings.TrimSpace(modelID)
-			if modelID == "" {
-				continue
-			}
-			if real, found := config.RealModelName(modelID); found {
-				toPreload = append(toPreload, real)
-			}
-		}
-		config.Hooks.OnStartup.Preload = toPreload
-	}
-
-	// Validate API keys (env macros already substituted at string level)
-	for i, apikey := range config.RequiredAPIKeys {
-		if apikey == "" {
-			return Config{}, fmt.Errorf("empty api key found in apiKeys")
-		}
-		if strings.Contains(apikey, " ") {
-			return Config{}, fmt.Errorf("api key cannot contain spaces: `%s`", apikey)
-		}
-		config.RequiredAPIKeys[i] = apikey
-	}
-
-	// Process peers with global macro substitution
-	for peerName, peerConfig := range config.Peers {
-		// Substitute global macros (LIFO order)
-		for i := len(config.Macros) - 1; i >= 0; i-- {
-			entry := config.Macros[i]
-			macroSlug := fmt.Sprintf("${%s}", entry.Name)
-			macroStr := fmt.Sprintf("%v", entry.Value)
-
-			peerConfig.ApiKey = strings.ReplaceAll(peerConfig.ApiKey, macroSlug, macroStr)
-			peerConfig.Filters.StripParams = strings.ReplaceAll(peerConfig.Filters.StripParams, macroSlug, macroStr)
-
-			// Substitute in setParams (type-preserving)
-			if len(peerConfig.Filters.SetParams) > 0 {
-				result, err := substituteMacroInValue(peerConfig.Filters.SetParams, entry.Name, entry.Value)
-				if err != nil {
-					return Config{}, fmt.Errorf("peers.%s.filters.setParams: %w", peerName, err)
-				}
-				peerConfig.Filters.SetParams = result.(map[string]any)
-			}
-		}
-
-		// Validate no unknown macros remain
-		if matches := macroPatternRegex.FindAllStringSubmatch(peerConfig.ApiKey, -1); len(matches) > 0 {
-			return Config{}, fmt.Errorf("peers.%s.apiKey: unknown macro '${%s}'", peerName, matches[0][1])
-		}
-		if matches := macroPatternRegex.FindAllStringSubmatch(peerConfig.Filters.StripParams, -1); len(matches) > 0 {
-			return Config{}, fmt.Errorf("peers.%s.filters.stripParams: unknown macro '${%s}'", peerName, matches[0][1])
-		}
-		if len(peerConfig.Filters.SetParams) > 0 {
-			if err := validateNestedForUnknownMacros(peerConfig.Filters.SetParams, fmt.Sprintf("peers.%s.filters.setParams", peerName)); err != nil {
-				return Config{}, err
-			}
-		}
-		config.Peers[peerName] = peerConfig
-	}
-
-	return config, nil
-}
-
 // rewrites the yaml to include a default group with any orphaned models
 func AddDefaultGroupToConfig(config Config) Config {
 
@@ -596,233 +255,3 @@ func AddDefaultGroupToConfig(config Config) Config {
 
 	return config
 }
-
-func SanitizeCommand(cmdStr string) ([]string, error) {
-	var cleanedLines []string
-	for _, line := range strings.Split(cmdStr, "\n") {
-		trimmed := strings.TrimSpace(line)
-		// Skip comment lines
-		if strings.HasPrefix(trimmed, "#") {
-			continue
-		}
-		// Handle trailing backslashes by replacing with space
-		if strings.HasSuffix(trimmed, "\\") {
-			cleanedLines = append(cleanedLines, strings.TrimSuffix(trimmed, "\\")+" ")
-		} else {
-			cleanedLines = append(cleanedLines, line)
-		}
-	}
-
-	// put it back together
-	cmdStr = strings.Join(cleanedLines, "\n")
-
-	// Split the command into arguments
-	var args []string
-	if runtime.GOOS == "windows" {
-		args = shlex.Windows.Split(cmdStr)
-	} else {
-		args = shlex.Posix.Split(cmdStr)
-	}
-
-	// Ensure the command is not empty
-	if len(args) == 0 {
-		return nil, fmt.Errorf("empty command")
-	}
-
-	return args, nil
-}
-
-func StripComments(cmdStr string) string {
-	var cleanedLines []string
-	for _, line := range strings.Split(cmdStr, "\n") {
-		trimmed := strings.TrimSpace(line)
-		// Skip comment lines
-		if strings.HasPrefix(trimmed, "#") {
-			continue
-		}
-		cleanedLines = append(cleanedLines, line)
-	}
-	return strings.Join(cleanedLines, "\n")
-}
-
-// validateMacro validates macro name and value constraints
-func validateMacro(name string, value any) error {
-	if len(name) >= 64 {
-		return fmt.Errorf("macro name '%s' exceeds maximum length of 63 characters", name)
-	}
-	if !macroNameRegex.MatchString(name) {
-		return fmt.Errorf("macro name '%s' contains invalid characters, must match pattern ^[a-zA-Z0-9_-]+$", name)
-	}
-
-	// Validate that value is a scalar type
-	switch v := value.(type) {
-	case string:
-		// Check for self-reference
-		macroSlug := fmt.Sprintf("${%s}", name)
-		if strings.Contains(v, macroSlug) {
-			return fmt.Errorf("macro '%s' contains self-reference", name)
-		}
-	case int, int8, int16, int32, int64, uint, uint8, uint16, uint32, uint64, float32, float64, bool:
-		// These types are allowed
-	default:
-		return fmt.Errorf("macro '%s' has invalid type %T, must be a scalar type (string, int, float, or bool)", name, value)
-	}
-
-	switch name {
-	case "PORT", "MODEL_ID":
-		return fmt.Errorf("macro name '%s' is reserved", name)
-	}
-
-	return nil
-}
-
-// validateNestedForUnknownMacros recursively checks for any remaining macro references in nested structures
-func validateNestedForUnknownMacros(value any, context string) error {
-	switch v := value.(type) {
-	case string:
-		matches := macroPatternRegex.FindAllStringSubmatch(v, -1)
-		for _, match := range matches {
-			macroName := match[1]
-			return fmt.Errorf("%s: unknown macro '${%s}'", context, macroName)
-		}
-		// Check for unsubstituted env macros
-		envMatches := envMacroRegex.FindAllStringSubmatch(v, -1)
-		for _, match := range envMatches {
-			varName := match[1]
-			return fmt.Errorf("%s: environment variable '%s' not set", context, varName)
-		}
-		return nil
-
-	case map[string]any:
-		for _, val := range v {
-			if err := validateNestedForUnknownMacros(val, context); err != nil {
-				return err
-			}
-		}
-		return nil
-
-	case []any:
-		for _, val := range v {
-			if err := validateNestedForUnknownMacros(val, context); err != nil {
-				return err
-			}
-		}
-		return nil
-
-	default:
-		// Scalar types don't contain macros
-		return nil
-	}
-}
-
-// substituteMacroInValue recursively substitutes a single macro in a value structure
-// This is called once per macro, allowing LIFO substitution order
-func substituteMacroInValue(value any, macroName string, macroValue any) (any, error) {
-	macroSlug := fmt.Sprintf("${%s}", macroName)
-	macroStr := fmt.Sprintf("%v", macroValue)
-
-	switch v := value.(type) {
-	case string:
-		// Check if this is a direct macro substitution
-		if v == macroSlug {
-			return macroValue, nil
-		}
-		// Handle string interpolation
-		if strings.Contains(v, macroSlug) {
-			return strings.ReplaceAll(v, macroSlug, macroStr), nil
-		}
-		return v, nil
-
-	case map[string]any:
-		// Recursively process map values
-		newMap := make(map[string]any)
-		for key, val := range v {
-			newVal, err := substituteMacroInValue(val, macroName, macroValue)
-			if err != nil {
-				return nil, err
-			}
-			newMap[key] = newVal
-		}
-		return newMap, nil
-
-	case []any:
-		// Recursively process slice elements
-		newSlice := make([]any, len(v))
-		for i, val := range v {
-			newVal, err := substituteMacroInValue(val, macroName, macroValue)
-			if err != nil {
-				return nil, err
-			}
-			newSlice[i] = newVal
-		}
-		return newSlice, nil
-
-	default:
-		// Return scalar types as-is
-		return value, nil
-	}
-}
-
-// substituteEnvMacros replaces ${env.VAR_NAME} with environment variable values.
-// Returns error if any referenced env var is not set or contains invalid characters.
-// Env macros inside YAML comments are ignored by unmarshalling the YAML first
-// (which strips comments) and only checking the comment-free version for macros.
-func substituteEnvMacros(s string) (string, error) {
-	// Unmarshal and remarshal to strip YAML comments
-	var raw any
-	if err := yaml.Unmarshal([]byte(s), &raw); err != nil {
-		// If YAML is invalid, fall back to scanning the original string
-		// so the user gets the env var error rather than a confusing YAML parse error
-		return substituteEnvMacrosInString(s, s)
-	}
-	clean, err := yaml.Marshal(raw)
-	if err != nil {
-		return substituteEnvMacrosInString(s, s)
-	}
-
-	return substituteEnvMacrosInString(s, string(clean))
-}
-
-// substituteEnvMacrosInString finds ${env.VAR} macros in scanStr and substitutes
-// them in target. This separation allows scanning comment-free YAML while
-// substituting in the original string.
-func substituteEnvMacrosInString(target, scanStr string) (string, error) {
-	result := target
-	matches := envMacroRegex.FindAllStringSubmatch(scanStr, -1)
-	for _, match := range matches {
-		fullMatch := match[0] // ${env.VAR_NAME}
-		varName := match[1]   // VAR_NAME
-
-		value, exists := os.LookupEnv(varName)
-		if !exists {
-			return "", fmt.Errorf("environment variable '%s' is not set", varName)
-		}
-
-		// Sanitize the value for safe YAML substitution
-		value, err := sanitizeEnvValueForYAML(value, varName)
-		if err != nil {
-			return "", err
-		}
-
-		result = strings.ReplaceAll(result, fullMatch, value)
-	}
-	return result, nil
-}
-
-// sanitizeEnvValueForYAML ensures an environment variable value is safe for YAML substitution.
-// It rejects values with characters that break YAML structure and escapes quotes/backslashes
-// for compatibility with double-quoted YAML strings.
-func sanitizeEnvValueForYAML(value, varName string) (string, error) {
-	// Reject values that would break YAML structure regardless of quoting context
-	if strings.ContainsAny(value, "\n\r\x00") {
-		return "", fmt.Errorf("environment variable '%s' contains newlines or null bytes which are not allowed in YAML substitution", varName)
-	}
-
-	// Escape backslashes and double quotes for safe use in double-quoted YAML strings.
-	// In unquoted contexts, these escapes appear literally (harmless for most use cases).
-	// In double-quoted contexts, they are interpreted correctly.
-	value = strings.ReplaceAll(value, `\`, `\\`)
-	value = strings.ReplaceAll(value, `"`, `\"`)
-
-	return value, nil
-}

internal/config/config_posix_test.go

@@ -173,6 +173,25 @@ groups:
 		IdleConn:       90,
 	}
 
+	expectedGroups := map[string]GroupConfig{
+		DEFAULT_GROUP_ID: {
+			Swap:      true,
+			Exclusive: true,
+			Members:   []string{"model1", "model3"},
+		},
+		"group1": {
+			Swap:      true,
+			Exclusive: false,
+			Members:   []string{"model2"},
+		},
+		"forever": {
+			Swap:       true,
+			Exclusive:  false,
+			Persistent: true,
+			Members:    []string{"model4"},
+		},
+	}
+
 	expected := Config{
 		LogLevel:      "info",
 		LogTimeFormat: "",
@@ -246,22 +265,19 @@ groups:
 			"m2":        "model2",
 			"mthree":    "model3",
 		},
-		Groups: map[string]GroupConfig{
-			DEFAULT_GROUP_ID: {
-				Swap:      true,
-				Exclusive: true,
-				Members:   []string{"model1", "model3"},
+		Groups: expectedGroups,
+		Upstream: UpstreamConfig{
+			IgnorePaths: DefaultUpstreamIgnorePaths(),
+		},
+		Routing: RoutingConfig{
+			Router: RouterConfig{
+				Use: "group",
+				Settings: RouterSettings{
+					Groups: expectedGroups,
+				},
 			},
-			"group1": {
-				Swap:      true,
-				Exclusive: false,
-				Members:   []string{"model2"},
-			},
-			"forever": {
-				Swap:       true,
-				Exclusive:  false,
-				Persistent: true,
-				Members:    []string{"model4"},
+			Scheduler: SchedulerConfig{
+				Use: "serial",
 			},
 		},
 	}

internal/config/config_schema_test.go

@@ -0,0 +1,60 @@
+package config
+
+import (
+	"encoding/json"
+	"os"
+	"testing"
+
+	"github.com/google/jsonschema-go/jsonschema"
+	"gopkg.in/yaml.v3"
+)
+
+// TestConfig_ExampleMatchesSchema validates that config.example.yaml conforms to
+// config-schema.json. Both files live at the repository root.
+func TestConfig_ExampleMatchesSchema(t *testing.T) {
+	const (
+		schemaPath  = "../../config-schema.json"
+		examplePath = "../../config.example.yaml"
+	)
+
+	schemaBytes, err := os.ReadFile(schemaPath)
+	if err != nil {
+		t.Fatalf("reading %s: %v", schemaPath, err)
+	}
+
+	var schema jsonschema.Schema
+	if err := json.Unmarshal(schemaBytes, &schema); err != nil {
+		t.Fatalf("unmarshalling schema: %v", err)
+	}
+
+	resolved, err := schema.Resolve(&jsonschema.ResolveOptions{
+		BaseURI: "https://github.com/mostlygeek/llama-swap/",
+	})
+	if err != nil {
+		t.Fatalf("resolving schema: %v", err)
+	}
+
+	exampleBytes, err := os.ReadFile(examplePath)
+	if err != nil {
+		t.Fatalf("reading %s: %v", examplePath, err)
+	}
+
+	// Convert YAML to a JSON-like value so numbers and keys match what the
+	// validator expects.
+	var yamlValue any
+	if err := yaml.Unmarshal(exampleBytes, &yamlValue); err != nil {
+		t.Fatalf("unmarshalling example yaml: %v", err)
+	}
+	jsonBytes, err := json.Marshal(yamlValue)
+	if err != nil {
+		t.Fatalf("converting example to json: %v", err)
+	}
+	var instance any
+	if err := json.Unmarshal(jsonBytes, &instance); err != nil {
+		t.Fatalf("unmarshalling example json: %v", err)
+	}
+
+	if err := resolved.Validate(instance); err != nil {
+		t.Fatalf("config.example.yaml does not match config-schema.json:\n%v", err)
+	}
+}

internal/config/config_test.go

@@ -777,22 +777,27 @@ func TestConfig_APIKeys_Invalid(t *testing.T) {
 		{
 			name:        "blank spaces only",
 			content:     `apiKeys: ["   "]`,
-			expectedErr: "api key cannot contain spaces: `   `",
+			expectedErr: "apiKeys[0]: api key cannot contain spaces",
 		},
 		{
 			name:        "contains leading space",
 			content:     `apiKeys: [" key123"]`,
-			expectedErr: "api key cannot contain spaces: ` key123`",
+			expectedErr: "apiKeys[0]: api key cannot contain spaces",
 		},
 		{
 			name:        "contains trailing space",
 			content:     `apiKeys: ["key123 "]`,
-			expectedErr: "api key cannot contain spaces: `key123 `",
+			expectedErr: "apiKeys[0]: api key cannot contain spaces",
 		},
 		{
 			name:        "contains middle space",
 			content:     `apiKeys: ["key 123"]`,
-			expectedErr: "api key cannot contain spaces: `key 123`",
+			expectedErr: "apiKeys[0]: api key cannot contain spaces",
+		},
+		{
+			name:        "space in second key reports correct index",
+			content:     `apiKeys: ["valid-key", "bad key"]`,
+			expectedErr: "apiKeys[1]: api key cannot contain spaces",
 		},
 		{
 			name:        "empty in list with valid keys",
@@ -1544,3 +1549,174 @@ peers:
 	assert.Equal(t, 1, peerConfig.Timeouts.ExpectContinue)
 	assert.Equal(t, 90, peerConfig.Timeouts.IdleConn)
 }
+
+// twoModels is a minimal models block reused by the routing tests below.
+const twoModels = `
+models:
+  gemma:
+    cmd: echo gemma
+    proxy: http://localhost:8080
+  qwen:
+    cmd: echo qwen
+    proxy: http://localhost:8081
+`
+
+func TestConfig_Routing_LegacyTopLevelGroups(t *testing.T) {
+	yaml := twoModels + `
+groups:
+  g1:
+    members: [gemma, qwen]
+`
+	cfg, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.NoError(t, err)
+	assert.Equal(t, "group", cfg.Routing.Router.Use)
+	// default group injected for orphaned models (none here) still leaves g1
+	assert.Contains(t, cfg.Routing.Router.Settings.Groups, "g1")
+	assert.Equal(t, "serial", cfg.Routing.Scheduler.Use)
+}
+
+func TestConfig_Routing_LegacyTopLevelMatrix(t *testing.T) {
+	yaml := twoModels + `
+matrix:
+  vars:
+    g: gemma
+    q: qwen
+  sets:
+    combo: "g | q"
+`
+	cfg, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.NoError(t, err)
+	assert.Equal(t, "matrix", cfg.Routing.Router.Use)
+	require.NotNil(t, cfg.Routing.Router.Settings.Matrix)
+	assert.Len(t, cfg.Routing.Router.Settings.Matrix.ExpandedSets, 2)
+}
+
+func TestConfig_Routing_RouterUseMatrix(t *testing.T) {
+	yaml := twoModels + `
+routing:
+  router:
+    use: matrix
+    settings:
+      matrix:
+        vars:
+          g: gemma
+          q: qwen
+        sets:
+          combo: "g | q"
+`
+	cfg, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.NoError(t, err)
+	assert.Equal(t, "matrix", cfg.Routing.Router.Use)
+	require.NotNil(t, cfg.Routing.Router.Settings.Matrix)
+	assert.Len(t, cfg.Routing.Router.Settings.Matrix.ExpandedSets, 2)
+}
+
+func TestConfig_Routing_RouterUseGroup(t *testing.T) {
+	yaml := twoModels + `
+routing:
+  router:
+    use: group
+    settings:
+      groups:
+        g1:
+          members: [gemma, qwen]
+`
+	cfg, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.NoError(t, err)
+	assert.Equal(t, "group", cfg.Routing.Router.Use)
+	assert.Contains(t, cfg.Routing.Router.Settings.Groups, "g1")
+}
+
+func TestConfig_Routing_DefaultsToGroup(t *testing.T) {
+	cfg, err := LoadConfigFromReader(strings.NewReader(twoModels))
+	require.NoError(t, err)
+	assert.Equal(t, "group", cfg.Routing.Router.Use)
+	assert.Equal(t, "serial", cfg.Routing.Scheduler.Use)
+}
+
+func TestConfig_Routing_LegacyAndRoutingConflict(t *testing.T) {
+	yaml := twoModels + `
+groups:
+  g1:
+    members: [gemma, qwen]
+routing:
+  router:
+    use: group
+`
+	_, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "migrate")
+}
+
+func TestConfig_Routing_RouterUseMatrixWithoutSettings(t *testing.T) {
+	yaml := twoModels + `
+routing:
+  router:
+    use: matrix
+`
+	_, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "routing.router.settings.matrix is not set")
+}
+
+// Both groups and matrix may be defined under routing.router.settings;
+// routing.router.use selects which one is active.
+func TestConfig_Routing_RouterSettingsBothGroupsAndMatrix(t *testing.T) {
+	yaml := twoModels + `
+routing:
+  router:
+    use: group
+    settings:
+      groups:
+        g1:
+          members: [gemma, qwen]
+      matrix:
+        sets:
+          s: "gemma"
+`
+	config, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.NoError(t, err)
+	// use: group means groups are active and matrix is ignored
+	assert.Equal(t, "group", config.Routing.Router.Use)
+	assert.Nil(t, config.Matrix)
+	assert.Contains(t, config.Groups, "g1")
+}
+
+func TestConfig_Routing_UnknownRouter(t *testing.T) {
+	yaml := twoModels + `
+routing:
+  router:
+    use: bogus
+`
+	_, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "unknown router")
+}
+
+func TestConfig_Routing_FifoPriorityUnknownModel(t *testing.T) {
+	yaml := twoModels + `
+routing:
+  scheduler:
+    settings:
+      fifo:
+        priority:
+          nope: 5
+`
+	_, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "unknown model")
+}
+
+func TestConfig_Routing_FifoPriorityKnownModel(t *testing.T) {
+	yaml := twoModels + `
+routing:
+  scheduler:
+    settings:
+      fifo:
+        priority:
+          gemma: 5
+`
+	cfg, err := LoadConfigFromReader(strings.NewReader(yaml))
+	require.NoError(t, err)
+	assert.Equal(t, 5, cfg.Routing.Scheduler.Settings.Fifo.Priority["gemma"])
+}

internal/config/config_windows_test.go

@@ -165,6 +165,25 @@ groups:
 		IdleConn:       90,
 	}
 
+	expectedGroups := map[string]GroupConfig{
+		DEFAULT_GROUP_ID: {
+			Swap:      true,
+			Exclusive: true,
+			Members:   []string{"model1", "model3"},
+		},
+		"group1": {
+			Swap:      true,
+			Exclusive: false,
+			Members:   []string{"model2"},
+		},
+		"forever": {
+			Swap:       true,
+			Exclusive:  false,
+			Persistent: true,
+			Members:    []string{"model4"},
+		},
+	}
+
 	expected := Config{
 		LogLevel:      "info",
 		LogTimeFormat: "",
@@ -235,22 +254,19 @@ groups:
 			"m2":        "model2",
 			"mthree":    "model3",
 		},
-		Groups: map[string]GroupConfig{
-			DEFAULT_GROUP_ID: {
-				Swap:      true,
-				Exclusive: true,
-				Members:   []string{"model1", "model3"},
+		Groups: expectedGroups,
+		Upstream: UpstreamConfig{
+			IgnorePaths: DefaultUpstreamIgnorePaths(),
+		},
+		Routing: RoutingConfig{
+			Router: RouterConfig{
+				Use: "group",
+				Settings: RouterSettings{
+					Groups: expectedGroups,
+				},
 			},
-			"group1": {
-				Swap:      true,
-				Exclusive: false,
-				Members:   []string{"model2"},
-			},
-			"forever": {
-				Swap:       true,
-				Exclusive:  false,
-				Persistent: true,
-				Members:    []string{"model4"},
+			Scheduler: SchedulerConfig{
+				Use: "serial",
 			},
 		},
 	}

internal/config/load.go

@@ -0,0 +1,441 @@
+package config
+
+import (
+	"fmt"
+	"io"
+	"net/url"
+	"sort"
+	"strings"
+	"time"
+
+	"gopkg.in/yaml.v3"
+)
+
+func LoadConfigFromReader(r io.Reader) (Config, error) {
+	data, err := io.ReadAll(r)
+	if err != nil {
+		return Config{}, err
+	}
+	yamlStr := string(data)
+
+	// Phase 1: Substitute all ${env.VAR} macros at string level
+	// This is safe because env values are simple strings without YAML formatting
+	yamlStr, err = substituteEnvMacros(yamlStr)
+	if err != nil {
+		return Config{}, err
+	}
+
+	// Unmarshal into full Config with defaults
+	config := Config{
+		HealthCheckTimeout: 120,
+		StartPort:          5800,
+		LogLevel:           "info",
+		LogTimeFormat:      "",
+		LogToStdout:        LogToStdoutProxy,
+		MetricsMaxInMemory: 1000,
+		CaptureBuffer:      5,
+		GlobalTTL:          0,
+	}
+	if err = yaml.Unmarshal([]byte(yamlStr), &config); err != nil {
+		return Config{}, err
+	}
+
+	if config.HealthCheckTimeout < 15 {
+		config.HealthCheckTimeout = 15
+	}
+
+	// Apply defaults for performance config when section is missing
+	if config.Performance.Every == 0 {
+		config.Performance.Every = 5 * time.Second
+	}
+	if err = config.Performance.Validate(); err != nil {
+		return Config{}, fmt.Errorf("performance: %w", err)
+	}
+
+	if config.StartPort < 1 {
+		return Config{}, fmt.Errorf("startPort must be greater than 1")
+	}
+
+	if config.GlobalTTL < 0 {
+		return Config{}, fmt.Errorf("globalTTL must be >= 0")
+	}
+
+	// Apply default for upstream.ignorePaths when not specified. The default
+	// matches common static-asset suffixes so they do not trigger a swap.
+	if len(config.Upstream.IgnorePaths) == 0 {
+		config.Upstream.IgnorePaths = DefaultUpstreamIgnorePaths()
+	}
+
+	switch config.LogToStdout {
+	case LogToStdoutProxy, LogToStdoutUpstream, LogToStdoutBoth, LogToStdoutNone:
+	default:
+		return Config{}, fmt.Errorf("logToStdout must be one of: proxy, upstream, both, none")
+	}
+
+	// Populate the aliases map
+	config.aliases = make(map[string]string)
+	for modelName, modelConfig := range config.Models {
+		for _, alias := range modelConfig.Aliases {
+			if _, found := config.aliases[alias]; found {
+				return Config{}, fmt.Errorf("duplicate alias %s found in model: %s", alias, modelName)
+			}
+			config.aliases[alias] = modelName
+		}
+	}
+
+	// Validate global macros
+	for _, macro := range config.Macros {
+		if err = validateMacro(macro.Name, macro.Value); err != nil {
+			return Config{}, err
+		}
+	}
+
+	// Get and sort all model IDs for consistent port assignment
+	modelIds := make([]string, 0, len(config.Models))
+	for modelId := range config.Models {
+		modelIds = append(modelIds, modelId)
+	}
+	sort.Strings(modelIds)
+
+	nextPort := config.StartPort
+	for _, modelId := range modelIds {
+		modelConfig := config.Models[modelId]
+		modelConfig.HealthCheckTimeout = config.HealthCheckTimeout
+
+		// Strip comments from command fields
+		modelConfig.Cmd = StripComments(modelConfig.Cmd)
+		modelConfig.CmdStop = StripComments(modelConfig.CmdStop)
+
+		// set model TTL to globalTTL it is the default value
+		if modelConfig.UnloadAfter == MODEL_CONFIG_DEFAULT_TTL {
+			modelConfig.UnloadAfter = config.GlobalTTL
+		}
+
+		if modelConfig.UnloadAfter < 0 {
+			return Config{}, fmt.Errorf("model %s: invalid TTL value %d", modelId, modelConfig.UnloadAfter)
+		}
+
+		// Validate model macros
+		for _, macro := range modelConfig.Macros {
+			if err = validateMacro(macro.Name, macro.Value); err != nil {
+				return Config{}, fmt.Errorf("model %s: %s", modelId, err.Error())
+			}
+		}
+
+		// Build merged macro list: MODEL_ID + global macros + model macros (model overrides global)
+		mergedMacros := make(MacroList, 0, len(config.Macros)+len(modelConfig.Macros)+1)
+		mergedMacros = append(mergedMacros, MacroEntry{Name: "MODEL_ID", Value: modelId})
+		mergedMacros = append(mergedMacros, config.Macros...)
+
+		// Add model macros (override globals with same name)
+		for _, entry := range modelConfig.Macros {
+			found := false
+			for i, existing := range mergedMacros {
+				if existing.Name == entry.Name {
+					mergedMacros[i] = entry
+					found = true
+					break
+				}
+			}
+			if !found {
+				mergedMacros = append(mergedMacros, entry)
+			}
+		}
+
+		// Substitute remaining macros in model fields (LIFO order)
+		for i := len(mergedMacros) - 1; i >= 0; i-- {
+			entry := mergedMacros[i]
+			macroSlug := fmt.Sprintf("${%s}", entry.Name)
+			macroStr := fmt.Sprintf("%v", entry.Value)
+
+			modelConfig.Cmd = strings.ReplaceAll(modelConfig.Cmd, macroSlug, macroStr)
+			modelConfig.CmdStop = strings.ReplaceAll(modelConfig.CmdStop, macroSlug, macroStr)
+			modelConfig.Proxy = strings.ReplaceAll(modelConfig.Proxy, macroSlug, macroStr)
+			modelConfig.CheckEndpoint = strings.ReplaceAll(modelConfig.CheckEndpoint, macroSlug, macroStr)
+			modelConfig.Filters.StripParams = strings.ReplaceAll(modelConfig.Filters.StripParams, macroSlug, macroStr)
+			modelConfig.Name = strings.ReplaceAll(modelConfig.Name, macroSlug, macroStr)
+			modelConfig.Description = strings.ReplaceAll(modelConfig.Description, macroSlug, macroStr)
+
+			// Substitute macros in SetParamsByID keys and values
+			if len(modelConfig.Filters.SetParamsByID) > 0 {
+				newSetParamsByID := make(map[string]map[string]any, len(modelConfig.Filters.SetParamsByID))
+				for key, paramMap := range modelConfig.Filters.SetParamsByID {
+					newKey := strings.ReplaceAll(key, macroSlug, macroStr)
+					newValAny, err := substituteMacroInValue(any(paramMap), entry.Name, entry.Value)
+					if err != nil {
+						return Config{}, fmt.Errorf("model %s filters.setParamsByID: %s", modelId, err.Error())
+					}
+					newParamMap, ok := newValAny.(map[string]any)
+					if !ok {
+						return Config{}, fmt.Errorf("model %s filters.setParamsByID: unexpected type after macro substitution", modelId)
+					}
+					newSetParamsByID[newKey] = newParamMap
+				}
+				modelConfig.Filters.SetParamsByID = newSetParamsByID
+			}
+
+			// Substitute in metadata (type-preserving)
+			if len(modelConfig.Metadata) > 0 {
+				result, err := substituteMacroInValue(modelConfig.Metadata, entry.Name, entry.Value)
+				if err != nil {
+					return Config{}, fmt.Errorf("model %s metadata: %s", modelId, err.Error())
+				}
+				modelConfig.Metadata = result.(map[string]any)
+			}
+		}
+
+		// Handle PORT macro - only allocate if cmd uses it
+		cmdHasPort := strings.Contains(modelConfig.Cmd, "${PORT}")
+		proxyHasPort := strings.Contains(modelConfig.Proxy, "${PORT}")
+		if cmdHasPort || proxyHasPort {
+			if !cmdHasPort && proxyHasPort {
+				return Config{}, fmt.Errorf("model %s: proxy uses ${PORT} but cmd does not - ${PORT} is only available when used in cmd", modelId)
+			}
+
+			macroSlug := "${PORT}"
+			macroStr := fmt.Sprintf("%v", nextPort)
+
+			modelConfig.Cmd = strings.ReplaceAll(modelConfig.Cmd, macroSlug, macroStr)
+			modelConfig.CmdStop = strings.ReplaceAll(modelConfig.CmdStop, macroSlug, macroStr)
+			modelConfig.Proxy = strings.ReplaceAll(modelConfig.Proxy, macroSlug, macroStr)
+			modelConfig.Name = strings.ReplaceAll(modelConfig.Name, macroSlug, macroStr)
+			modelConfig.Description = strings.ReplaceAll(modelConfig.Description, macroSlug, macroStr)
+
+			if len(modelConfig.Metadata) > 0 {
+				result, err := substituteMacroInValue(modelConfig.Metadata, "PORT", nextPort)
+				if err != nil {
+					return Config{}, fmt.Errorf("model %s metadata: %s", modelId, err.Error())
+				}
+				modelConfig.Metadata = result.(map[string]any)
+			}
+
+			nextPort++
+		}
+
+		// Validate no unknown macros remain
+		fieldMap := map[string]string{
+			"cmd":                 modelConfig.Cmd,
+			"cmdStop":             modelConfig.CmdStop,
+			"proxy":               modelConfig.Proxy,
+			"checkEndpoint":       modelConfig.CheckEndpoint,
+			"filters.stripParams": modelConfig.Filters.StripParams,
+			"name":                modelConfig.Name,
+			"description":         modelConfig.Description,
+		}
+
+		for fieldName, fieldValue := range fieldMap {
+			matches := macroPatternRegex.FindAllStringSubmatch(fieldValue, -1)
+			for _, match := range matches {
+				macroName := match[1]
+				if macroName == "PID" && fieldName == "cmdStop" {
+					continue // replaced at runtime
+				}
+				if macroName == "PORT" || macroName == "MODEL_ID" {
+					return Config{}, fmt.Errorf("macro '${%s}' should have been substituted in %s.%s", macroName, modelId, fieldName)
+				}
+				return Config{}, fmt.Errorf("unknown macro '${%s}' found in %s.%s", macroName, modelId, fieldName)
+			}
+		}
+
+		if len(modelConfig.Metadata) > 0 {
+			if err := validateNestedForUnknownMacros(modelConfig.Metadata, fmt.Sprintf("model %s metadata", modelId)); err != nil {
+				return Config{}, err
+			}
+		}
+
+		if err = modelConfig.Capabilities.Validate(); err != nil {
+			return Config{}, fmt.Errorf("model %s: %w", modelId, err)
+		}
+
+		// Validate SetParamsByID keys and values
+		for key, paramMap := range modelConfig.Filters.SetParamsByID {
+			if matches := macroPatternRegex.FindAllStringSubmatch(key, -1); len(matches) > 0 {
+				return Config{}, fmt.Errorf("unknown macro '${%s}' found in model %s filters.setParamsByID key", matches[0][1], modelId)
+			}
+			if err := validateNestedForUnknownMacros(any(paramMap), fmt.Sprintf("model %s filters.setParamsByID[%s]", modelId, key)); err != nil {
+				return Config{}, err
+			}
+		}
+
+		// Auto-register setParamsByID keys as aliases (skip the model's own ID)
+		for key := range modelConfig.Filters.SetParamsByID {
+			if key == modelId {
+				continue
+			}
+			if _, exists := config.Models[key]; exists {
+				return Config{}, fmt.Errorf("model %s filters.setParamsByID: key '%s' conflicts with an existing model ID", modelId, key)
+			}
+			if existingModel, exists := config.aliases[key]; exists {
+				if existingModel != modelId {
+					return Config{}, fmt.Errorf("duplicate alias '%s' in model %s filters.setParamsByID, already used by model %s", key, modelId, existingModel)
+				}
+				continue // already registered as explicit alias for this model
+			}
+			config.aliases[key] = modelId
+			modelConfig.Aliases = append(modelConfig.Aliases, key)
+		}
+
+		if _, err := url.Parse(modelConfig.Proxy); err != nil {
+			return Config{}, fmt.Errorf("model %s: invalid proxy URL: %w", modelId, err)
+		}
+
+		if modelConfig.SendLoadingState == nil {
+			v := config.SendLoadingState
+			modelConfig.SendLoadingState = &v
+		}
+
+		config.Models[modelId] = modelConfig
+	}
+
+	// Normalize routing config. The legacy top-level `matrix`/`groups` keys and
+	// the new `routing.router` block are mutually exclusive: a config may use
+	// either style, never both.
+	hasTopLevel := config.Matrix != nil || len(config.Groups) > 0
+	rtr := config.Routing.Router
+	hasRouting := rtr.Use != "" || rtr.Settings.Matrix != nil || len(rtr.Settings.Groups) > 0
+
+	if hasTopLevel && hasRouting {
+		return Config{}, fmt.Errorf("config uses both the legacy top-level 'matrix'/'groups' keys and the new 'routing.router' block; please migrate the top-level keys into 'routing.router' and remove them")
+	}
+
+	if !hasTopLevel {
+		// Both groups and matrix may be defined under routing.router.settings;
+		// routing.router.use selects which one is active, so there is no conflict.
+		rs := config.Routing.Router.Settings
+		switch config.Routing.Router.Use {
+		case "matrix":
+			if rs.Matrix == nil {
+				return Config{}, fmt.Errorf("routing.router.use is 'matrix' but routing.router.settings.matrix is not set")
+			}
+			config.Matrix = rs.Matrix
+		case "group", "":
+			config.Groups = rs.Groups
+		default:
+			return Config{}, fmt.Errorf("routing.router.use: unknown router %q (valid: group, matrix)", config.Routing.Router.Use)
+		}
+	}
+
+	// groups XOR matrix
+	if config.Matrix != nil && len(config.Groups) > 0 {
+		return Config{}, fmt.Errorf("config cannot use both 'groups' and 'matrix'")
+	}
+
+	if config.Matrix != nil {
+		expandedSets, err := ValidateMatrix(*config.Matrix, config.Models)
+		if err != nil {
+			return Config{}, fmt.Errorf("matrix: %w", err)
+		}
+		config.Matrix.ExpandedSets = expandedSets
+	} else {
+		config = AddDefaultGroupToConfig(config)
+
+		// Validate group members
+		memberUsage := make(map[string]string)
+		for groupID, groupConfig := range config.Groups {
+			prevSet := make(map[string]bool)
+			for _, member := range groupConfig.Members {
+				if _, found := prevSet[member]; found {
+					return Config{}, fmt.Errorf("duplicate model member %s found in group: %s", member, groupID)
+				}
+				prevSet[member] = true
+
+				if existingGroup, exists := memberUsage[member]; exists {
+					return Config{}, fmt.Errorf("model member %s is used in multiple groups: %s and %s", member, existingGroup, groupID)
+				}
+				memberUsage[member] = groupID
+			}
+		}
+	}
+
+	// Build the canonical Config.Routing from the effective result. Both legacy
+	// and new-style configs converge here. The Matrix pointer is shared so
+	// ExpandedSets stays in one place.
+	if config.Matrix != nil {
+		config.Routing.Router.Use = "matrix"
+	} else {
+		config.Routing.Router.Use = "group"
+	}
+	config.Routing.Router.Settings.Matrix = config.Matrix
+	config.Routing.Router.Settings.Groups = config.Groups
+
+	// This fork defaults to the "serial" scheduler: one model loaded at a time,
+	// requests served in strict arrival order. Set use: fifo for the upstream
+	// throughput-oriented behavior that batches same-model requests.
+	if config.Routing.Scheduler.Use == "" {
+		config.Routing.Scheduler.Use = "serial"
+	}
+	switch config.Routing.Scheduler.Use {
+	case "fifo", "serial":
+	default:
+		return Config{}, fmt.Errorf("routing.scheduler.use: unknown scheduler %q (valid: fifo, serial)", config.Routing.Scheduler.Use)
+	}
+	for modelID := range config.Routing.Scheduler.Settings.Fifo.Priority {
+		if _, found := config.RealModelName(modelID); !found {
+			return Config{}, fmt.Errorf("routing.scheduler.settings.fifo.priority references unknown model %q", modelID)
+		}
+	}
+
+	// Clean up hooks preload
+	if len(config.Hooks.OnStartup.Preload) > 0 {
+		var toPreload []string
+		for _, modelID := range config.Hooks.OnStartup.Preload {
+			modelID = strings.TrimSpace(modelID)
+			if modelID == "" {
+				continue
+			}
+			if real, found := config.RealModelName(modelID); found {
+				toPreload = append(toPreload, real)
+			}
+		}
+		config.Hooks.OnStartup.Preload = toPreload
+	}
+
+	// Validate API keys (env macros already substituted at string level)
+	for i, apikey := range config.RequiredAPIKeys {
+		if apikey == "" {
+			return Config{}, fmt.Errorf("empty api key found in apiKeys")
+		}
+		if strings.Contains(apikey, " ") {
+			return Config{}, fmt.Errorf("apiKeys[%d]: api key cannot contain spaces", i)
+		}
+		config.RequiredAPIKeys[i] = apikey
+	}
+
+	// Process peers with global macro substitution
+	for peerName, peerConfig := range config.Peers {
+		// Substitute global macros (LIFO order)
+		for i := len(config.Macros) - 1; i >= 0; i-- {
+			entry := config.Macros[i]
+			macroSlug := fmt.Sprintf("${%s}", entry.Name)
+			macroStr := fmt.Sprintf("%v", entry.Value)
+
+			peerConfig.ApiKey = strings.ReplaceAll(peerConfig.ApiKey, macroSlug, macroStr)
+			peerConfig.Filters.StripParams = strings.ReplaceAll(peerConfig.Filters.StripParams, macroSlug, macroStr)
+
+			// Substitute in setParams (type-preserving)
+			if len(peerConfig.Filters.SetParams) > 0 {
+				result, err := substituteMacroInValue(peerConfig.Filters.SetParams, entry.Name, entry.Value)
+				if err != nil {
+					return Config{}, fmt.Errorf("peers.%s.filters.setParams: %w", peerName, err)
+				}
+				peerConfig.Filters.SetParams = result.(map[string]any)
+			}
+		}
+
+		// Validate no unknown macros remain
+		if matches := macroPatternRegex.FindAllStringSubmatch(peerConfig.ApiKey, -1); len(matches) > 0 {
+			return Config{}, fmt.Errorf("peers.%s.apiKey: unknown macro '${%s}'", peerName, matches[0][1])
+		}
+		if matches := macroPatternRegex.FindAllStringSubmatch(peerConfig.Filters.StripParams, -1); len(matches) > 0 {
+			return Config{}, fmt.Errorf("peers.%s.filters.stripParams: unknown macro '${%s}'", peerName, matches[0][1])
+		}
+		if len(peerConfig.Filters.SetParams) > 0 {
+			if err := validateNestedForUnknownMacros(peerConfig.Filters.SetParams, fmt.Sprintf("peers.%s.filters.setParams", peerName)); err != nil {
+				return Config{}, err
+			}
+		}
+		config.Peers[peerName] = peerConfig
+	}
+
+	return config, nil
+}

internal/config/macros.go

@@ -0,0 +1,198 @@
+package config
+
+import (
+	"fmt"
+	"os"
+	"regexp"
+	"strings"
+
+	"gopkg.in/yaml.v3"
+)
+
+var (
+	macroNameRegex    = regexp.MustCompile(`^[a-zA-Z0-9_-]+$`)
+	macroPatternRegex = regexp.MustCompile(`\$\{([a-zA-Z0-9_-]+)\}`)
+	envMacroRegex     = regexp.MustCompile(`\$\{env\.([a-zA-Z_][a-zA-Z0-9_]*)\}`)
+)
+
+// validateMacro validates macro name and value constraints
+func validateMacro(name string, value any) error {
+	if len(name) >= 64 {
+		return fmt.Errorf("macro name '%s' exceeds maximum length of 63 characters", name)
+	}
+	if !macroNameRegex.MatchString(name) {
+		return fmt.Errorf("macro name '%s' contains invalid characters, must match pattern ^[a-zA-Z0-9_-]+$", name)
+	}
+
+	// Validate that value is a scalar type
+	switch v := value.(type) {
+	case string:
+		// Check for self-reference
+		macroSlug := fmt.Sprintf("${%s}", name)
+		if strings.Contains(v, macroSlug) {
+			return fmt.Errorf("macro '%s' contains self-reference", name)
+		}
+	case int, int8, int16, int32, int64, uint, uint8, uint16, uint32, uint64, float32, float64, bool:
+		// These types are allowed
+	default:
+		return fmt.Errorf("macro '%s' has invalid type %T, must be a scalar type (string, int, float, or bool)", name, value)
+	}
+
+	switch name {
+	case "PORT", "MODEL_ID":
+		return fmt.Errorf("macro name '%s' is reserved", name)
+	}
+
+	return nil
+}
+
+// validateNestedForUnknownMacros recursively checks for any remaining macro references in nested structures
+func validateNestedForUnknownMacros(value any, context string) error {
+	switch v := value.(type) {
+	case string:
+		matches := macroPatternRegex.FindAllStringSubmatch(v, -1)
+		for _, match := range matches {
+			macroName := match[1]
+			return fmt.Errorf("%s: unknown macro '${%s}'", context, macroName)
+		}
+		// Check for unsubstituted env macros
+		envMatches := envMacroRegex.FindAllStringSubmatch(v, -1)
+		for _, match := range envMatches {
+			varName := match[1]
+			return fmt.Errorf("%s: environment variable '%s' not set", context, varName)
+		}
+		return nil
+
+	case map[string]any:
+		for _, val := range v {
+			if err := validateNestedForUnknownMacros(val, context); err != nil {
+				return err
+			}
+		}
+		return nil
+
+	case []any:
+		for _, val := range v {
+			if err := validateNestedForUnknownMacros(val, context); err != nil {
+				return err
+			}
+		}
+		return nil
+
+	default:
+		// Scalar types don't contain macros
+		return nil
+	}
+}
+
+// substituteMacroInValue recursively substitutes a single macro in a value structure
+// This is called once per macro, allowing LIFO substitution order
+func substituteMacroInValue(value any, macroName string, macroValue any) (any, error) {
+	macroSlug := fmt.Sprintf("${%s}", macroName)
+	macroStr := fmt.Sprintf("%v", macroValue)
+
+	switch v := value.(type) {
+	case string:
+		// Check if this is a direct macro substitution
+		if v == macroSlug {
+			return macroValue, nil
+		}
+		// Handle string interpolation
+		if strings.Contains(v, macroSlug) {
+			return strings.ReplaceAll(v, macroSlug, macroStr), nil
+		}
+		return v, nil
+
+	case map[string]any:
+		// Recursively process map values
+		newMap := make(map[string]any)
+		for key, val := range v {
+			newVal, err := substituteMacroInValue(val, macroName, macroValue)
+			if err != nil {
+				return nil, err
+			}
+			newMap[key] = newVal
+		}
+		return newMap, nil
+
+	case []any:
+		// Recursively process slice elements
+		newSlice := make([]any, len(v))
+		for i, val := range v {
+			newVal, err := substituteMacroInValue(val, macroName, macroValue)
+			if err != nil {
+				return nil, err
+			}
+			newSlice[i] = newVal
+		}
+		return newSlice, nil
+
+	default:
+		// Return scalar types as-is
+		return value, nil
+	}
+}
+
+// substituteEnvMacros replaces ${env.VAR_NAME} with environment variable values.
+// Returns error if any referenced env var is not set or contains invalid characters.
+// Env macros inside YAML comments are ignored by unmarshalling the YAML first
+// (which strips comments) and only checking the comment-free version for macros.
+func substituteEnvMacros(s string) (string, error) {
+	// Unmarshal and remarshal to strip YAML comments
+	var raw any
+	if err := yaml.Unmarshal([]byte(s), &raw); err != nil {
+		// If YAML is invalid, fall back to scanning the original string
+		// so the user gets the env var error rather than a confusing YAML parse error
+		return substituteEnvMacrosInString(s, s)
+	}
+	clean, err := yaml.Marshal(raw)
+	if err != nil {
+		return substituteEnvMacrosInString(s, s)
+	}
+
+	return substituteEnvMacrosInString(s, string(clean))
+}
+
+// substituteEnvMacrosInString finds ${env.VAR} macros in scanStr and substitutes
+// them in target. This separation allows scanning comment-free YAML while
+// substituting in the original string.
+func substituteEnvMacrosInString(target, scanStr string) (string, error) {
+	result := target
+	matches := envMacroRegex.FindAllStringSubmatch(scanStr, -1)
+	for _, match := range matches {
+		fullMatch := match[0] // ${env.VAR_NAME}
+		varName := match[1]   // VAR_NAME
+
+		value, exists := os.LookupEnv(varName)
+		if !exists {
+			return "", fmt.Errorf("environment variable '%s' is not set", varName)
+		}
+
+		// Sanitize the value for safe YAML substitution
+		value, err := sanitizeEnvValueForYAML(value, varName)
+		if err != nil {
+			return "", err
+		}
+
+		result = strings.ReplaceAll(result, fullMatch, value)
+	}
+	return result, nil
+}
+
+// sanitizeEnvValueForYAML ensures an environment variable value is safe for YAML substitution.
+// It rejects values with characters that break YAML structure and escapes quotes/backslashes
+// for compatibility with double-quoted YAML strings.
+func sanitizeEnvValueForYAML(value, varName string) (string, error) {
+	// Reject values that would break YAML structure regardless of quoting context
+	if strings.ContainsAny(value, "\n\r\x00") {
+		return "", fmt.Errorf("environment variable '%s' contains newlines or null bytes which are not allowed in YAML substitution", varName)
+	}
+
+	// Escape backslashes and double quotes for safe use in double-quoted YAML strings.
+	// In unquoted contexts, these escapes appear literally (harmless for most use cases).
+	// In double-quoted contexts, they are interpreted correctly.
+	value = strings.ReplaceAll(value, `\`, `\\`)
+	value = strings.ReplaceAll(value, `"`, `\"`)
+
+	return value, nil
+}

internal/config/matrix.go

@@ -15,6 +15,9 @@ type MatrixConfig struct {
 	Var        map[string]string `yaml:"vars"`
 	EvictCosts map[string]int    `yaml:"evict_costs"`
 	Sets       OrderedSets       `yaml:"sets"`
+
+	// populated by ValidateMatrix; not settable from yaml
+	ExpandedSets []ExpandedSet `yaml:"-"`
 }
 
 // SetEntry is a single named set with its DSL expression.

internal/config/matrix_test.go

@@ -289,7 +289,9 @@ matrix:
 	cfg, err := LoadConfigFromReader(strings.NewReader(yaml))
 	require.NoError(t, err)
 	assert.NotNil(t, cfg.Matrix)
-	assert.Len(t, cfg.ExpandedSets, 2)
+	assert.Len(t, cfg.Matrix.ExpandedSets, 2)
+	assert.Equal(t, "matrix", cfg.Routing.Router.Use)
+	assert.Len(t, cfg.Routing.Router.Settings.Matrix.ExpandedSets, 2)
 	// Groups should be empty when matrix is used
 	assert.Empty(t, cfg.Groups)
 }

internal/config/merge.go

@@ -0,0 +1,300 @@
+package config
+
+import (
+	"fmt"
+	"os"
+	"path/filepath"
+	"sort"
+	"strings"
+
+	"gopkg.in/yaml.v3"
+)
+
+// identityMapPaths is the set of dotted paths whose direct children are
+// identity-keyed maps. A child key present in two sources is a hard error;
+// such keys name discrete entities (a model, a group, a peer, etc.) and a
+// duplicate means the user has split one entity across files by mistake.
+var identityMapPaths = map[string]bool{
+	"models":                         true,
+	"groups":                         true,
+	"profiles":                       true,
+	"peers":                          true,
+	"matrix":                         true,
+	"routing.router.settings.groups": true,
+	"routing.router.settings.matrix": true,
+}
+
+// LoadConfigSources loads and merges configuration from -config (optional)
+// and -config-dir (optional). At least one must be provided. The -config file
+// is loaded first; *.yml/*.yaml files directly under -config-dir are then
+// merged in sorted filename order. The merged document is passed through the
+// existing LoadConfigFromReader pipeline unchanged.
+func LoadConfigSources(configPath, configDir string) (Config, error) {
+	if configPath == "" && configDir == "" {
+		return Config{}, fmt.Errorf("at least one of -config or -config-dir must be provided")
+	}
+
+	var sourcePaths []string
+
+	if configPath != "" {
+		sourcePaths = append(sourcePaths, configPath)
+	}
+
+	if configDir != "" {
+		dirFiles, err := listYAMLFiles(configDir)
+		if err != nil {
+			return Config{}, fmt.Errorf("-config-dir %s: %w", configDir, err)
+		}
+
+		if configPath != "" {
+			absConfig, err := filepath.Abs(configPath)
+			if err != nil {
+				return Config{}, fmt.Errorf("failed to resolve -config path: %w", err)
+			}
+			for _, f := range dirFiles {
+				absF, err := filepath.Abs(f)
+				if err != nil {
+					return Config{}, fmt.Errorf("failed to resolve config dir file %s: %w", f, err)
+				}
+				if absConfig == absF {
+					return Config{}, fmt.Errorf("-config path %s is also present in -config-dir %s; remove it from one", configPath, configDir)
+				}
+			}
+		}
+
+		sourcePaths = append(sourcePaths, dirFiles...)
+	}
+
+	if len(sourcePaths) == 0 {
+		return Config{}, fmt.Errorf("no configuration sources found")
+	}
+
+	var merged *yaml.Node
+	for _, p := range sourcePaths {
+		node, err := parseSource(p)
+		if err != nil {
+			return Config{}, err
+		}
+		if node == nil {
+			continue // empty file
+		}
+		if merged == nil {
+			merged = node
+			continue
+		}
+		if err := mergeNodes(merged, node, "", p); err != nil {
+			return Config{}, err
+		}
+	}
+
+	if merged == nil {
+		// All sources were empty; run the pipeline on empty input so defaults
+		// and validation still apply (e.g. startPort, performance defaults).
+		return LoadConfigFromReader(strings.NewReader(""))
+	}
+
+	out, err := yaml.Marshal(merged)
+	if err != nil {
+		return Config{}, fmt.Errorf("failed to marshal merged config: %w", err)
+	}
+	return LoadConfigFromReader(strings.NewReader(string(out)))
+}
+
+// listYAMLFiles returns the top-level *.yml and *.yaml files in dir, sorted by
+// filename for deterministic merge order. Subdirectories are not traversed.
+func listYAMLFiles(dir string) ([]string, error) {
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		return nil, err
+	}
+	var files []string
+	for _, e := range entries {
+		if e.IsDir() {
+			continue
+		}
+		name := e.Name()
+		if !strings.HasSuffix(name, ".yml") && !strings.HasSuffix(name, ".yaml") {
+			continue
+		}
+		files = append(files, filepath.Join(dir, name))
+	}
+	sort.Strings(files)
+	return files, nil
+}
+
+// parseSource reads and parses one YAML config file into a root mapping node.
+// Returns a nil node (no error) when the file is empty or contains only
+// comments.
+//
+// Env macros (${env.VAR}) are substituted at the string level before YAML
+// parsing so that flow-style constructs like [${env.API_KEY}] parse
+// correctly — the brace would otherwise be interpreted as a flow mapping.
+func parseSource(path string) (*yaml.Node, error) {
+	data, err := os.ReadFile(path)
+	if err != nil {
+		return nil, fmt.Errorf("failed to read config %s: %w", path, err)
+	}
+	yamlStr, err := substituteEnvMacros(string(data))
+	if err != nil {
+		return nil, fmt.Errorf("config %s: %w", path, err)
+	}
+	var doc yaml.Node
+	if err := yaml.Unmarshal([]byte(yamlStr), &doc); err != nil {
+		return nil, fmt.Errorf("failed to parse config %s: %w", path, err)
+	}
+	// yaml.Unmarshal into a yaml.Node yields a DocumentNode whose Content[0]
+	// is the actual root. Unwrap it so callers see the real top-level node.
+	root := &doc
+	if root.Kind == yaml.DocumentNode && len(root.Content) > 0 {
+		root = root.Content[0]
+	}
+	if root.Kind == 0 || root.Content == nil {
+		return nil, nil
+	}
+	if root.Kind != yaml.MappingNode {
+		return nil, fmt.Errorf("config %s: top-level YAML must be a mapping", path)
+	}
+	return root, nil
+}
+
+// mergeNodes merges src into dst (both MappingNodes) in place. Keys present in
+// only one side are kept; shared keys are merged recursively under the rules
+// in mergeValue. srcPath is included in error messages to identify the file
+// that introduced the conflict.
+func mergeNodes(dst, src *yaml.Node, path, srcPath string) error {
+	srcIdx := indexMapping(src)
+
+	// First pass: merge shared keys in place.
+	for i := 0; i+1 < len(dst.Content); i += 2 {
+		keyNode := dst.Content[i]
+		dstVal := dst.Content[i+1]
+		key := keyNode.Value
+
+		srcVal, ok := srcIdx[key]
+		if !ok {
+			continue // dst-only key, keep as-is
+		}
+
+		childPath := joinPath(path, key)
+
+		if identityMapPaths[childPath] {
+			// Identity-keyed map: each child key names a discrete entity
+			// (a model, group, peer, ...). A shared child key is a hard
+			// error; src-only children are appended in the second pass.
+			if err := mergeIdentityMap(dstVal, srcVal, childPath, key, srcPath); err != nil {
+				return err
+			}
+			continue
+		}
+
+		if err := mergeValue(dstVal, srcVal, childPath, srcPath); err != nil {
+			return err
+		}
+	}
+
+	// Second pass: append src-only keys.
+	dstIdx := indexMapping(dst)
+	for i := 0; i+1 < len(src.Content); i += 2 {
+		keyNode := src.Content[i]
+		srcVal := src.Content[i+1]
+		key := keyNode.Value
+
+		if _, ok := dstIdx[key]; ok {
+			continue // already merged above
+		}
+		keyCopy := *keyNode
+		valCopy := *srcVal
+		dst.Content = append(dst.Content, &keyCopy, &valCopy)
+	}
+
+	return nil
+}
+
+// mergeIdentityMap merges two identity-keyed mapping nodes (e.g. `models`,
+// `groups`, `peers`). Any child key present in both sides is a duplicate
+// entity and produces an error naming the conflicting key and source file.
+// src-only keys are appended to dst.
+func mergeIdentityMap(dst, src *yaml.Node, path, mapName, srcPath string) error {
+	if dst.Kind != yaml.MappingNode || src.Kind != yaml.MappingNode {
+		return fmt.Errorf("conflict at %q: expected a mapping, introduced by %s", path, srcPath)
+	}
+	dstIdx := indexMapping(dst)
+	for i := 0; i+1 < len(src.Content); i += 2 {
+		keyNode := src.Content[i]
+		srcVal := src.Content[i+1]
+		key := keyNode.Value
+		if _, dup := dstIdx[key]; dup {
+			return fmt.Errorf("duplicate %s %q found in %s (already defined in another config source)", mapName, key, srcPath)
+		}
+		keyCopy := *keyNode
+		valCopy := *srcVal
+		dst.Content = append(dst.Content, &keyCopy, &valCopy)
+	}
+	return nil
+}
+
+// mergeValue merges srcVal into dstVal (both pointing into the parent's
+// Content slice). Mapping+Mapping recurses; Sequence+Sequence concatenates;
+// Scalar+Scalar errors on value mismatch; null on either side yields to the
+// non-null side.
+func mergeValue(dstVal, srcVal *yaml.Node, path, srcPath string) error {
+	switch {
+	case dstVal.Kind == yaml.MappingNode && srcVal.Kind == yaml.MappingNode:
+		return mergeNodes(dstVal, srcVal, path, srcPath)
+
+	case dstVal.Kind == yaml.SequenceNode && srcVal.Kind == yaml.SequenceNode:
+		dstVal.Content = append(dstVal.Content, srcVal.Content...)
+		return nil
+
+	case dstVal.Kind == yaml.ScalarNode && srcVal.Kind == yaml.ScalarNode:
+		if isNullScalar(dstVal) {
+			*dstVal = *srcVal
+			return nil
+		}
+		if isNullScalar(srcVal) {
+			return nil
+		}
+		if dstVal.Value == srcVal.Value && dstVal.Tag == srcVal.Tag {
+			return nil
+		}
+		return fmt.Errorf("conflict at %q: %s sets a different value than a previous source", path, srcPath)
+
+	case isNull(dstVal):
+		*dstVal = *srcVal
+		return nil
+
+	case isNull(srcVal):
+		return nil
+
+	default:
+		return fmt.Errorf("conflict at %q: incompatible YAML node kinds (kind %d vs %d) introduced by %s", path, dstVal.Kind, srcVal.Kind, srcPath)
+	}
+}
+
+// isNull reports whether n represents a YAML null (empty or !!null).
+func isNull(n *yaml.Node) bool {
+	if n == nil || n.Kind == 0 {
+		return true
+	}
+	return isNullScalar(n)
+}
+
+func isNullScalar(n *yaml.Node) bool {
+	return n.Kind == yaml.ScalarNode && (n.Tag == "!!null" || n.Tag == "") && n.Value == ""
+}
+
+// indexMapping builds a key -> value-node index for a mapping node.
+func indexMapping(n *yaml.Node) map[string]*yaml.Node {
+	idx := make(map[string]*yaml.Node, len(n.Content)/2)
+	for i := 0; i+1 < len(n.Content); i += 2 {
+		idx[n.Content[i].Value] = n.Content[i+1]
+	}
+	return idx
+}
+
+func joinPath(parent, key string) string {
+	if parent == "" {
+		return key
+	}
+	return parent + "." + key
+}

internal/config/merge_test.go

@@ -0,0 +1,304 @@
+package config
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+// writeYAML writes content to a file named name inside dir. Returns the full
+// path of the written file.
+func writeYAML(t *testing.T, dir, name, content string) string {
+	t.Helper()
+	p := filepath.Join(dir, name)
+	require.NoError(t, os.MkdirAll(filepath.Dir(p), 0o755))
+	require.NoError(t, os.WriteFile(p, []byte(content), 0o644))
+	return p
+}
+
+// modelCfg builds a single-model YAML snippet indented for nesting under a
+// `models:` key. The proxy uses a fixed port so tests don't depend on
+// ${PORT} allocation.
+func modelCfg(id, cmd string) string {
+	return "  " + id + ":\n    cmd: " + cmd + "\n    proxy: \"http://localhost:9999\"\n"
+}
+
+func TestLoadConfigSources_NeitherProvided(t *testing.T) {
+	_, err := LoadConfigSources("", "")
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "at least one of -config or -config-dir")
+}
+
+func TestLoadConfigSources_ConfigOnly(t *testing.T) {
+	dir := t.TempDir()
+	cfgPath := writeYAML(t, dir, "config.yaml", `
+models:
+`+modelCfg("model1", "echo hi")+`
+groups:
+  group1:
+    members: ["model1"]
+`)
+	cfg, err := LoadConfigSources(cfgPath, "")
+	require.NoError(t, err)
+	_, id, ok := cfg.FindConfig("model1")
+	require.True(t, ok)
+	assert.Equal(t, "model1", id)
+}
+
+func TestLoadConfigSources_DirOnly(t *testing.T) {
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", "models:\n"+modelCfg("alpha", "echo a"))
+	writeYAML(t, dir, "b.yaml", "models:\n"+modelCfg("beta", "echo b"))
+
+	cfg, err := LoadConfigSources("", dir)
+	require.NoError(t, err)
+	for _, want := range []string{"alpha", "beta"} {
+		_, _, ok := cfg.FindConfig(want)
+		assert.True(t, ok, "model %s should be present", want)
+	}
+}
+
+func TestLoadConfigSources_ConfigPlusDirAdditive(t *testing.T) {
+	// -config lives outside -config-dir; both contribute models additively.
+	dir := t.TempDir()
+	cfgPath := writeYAML(t, dir, "config.yaml", "models:\n"+modelCfg("base", "echo base"))
+	cfgDir := t.TempDir()
+	writeYAML(t, cfgDir, "extra.yaml", "models:\n"+modelCfg("ext", "echo ext"))
+
+	cfg, err := LoadConfigSources(cfgPath, cfgDir)
+	require.NoError(t, err)
+	for _, want := range []string{"base", "ext"} {
+		_, _, ok := cfg.FindConfig(want)
+		assert.True(t, ok, "model %s should be present after merge", want)
+	}
+}
+
+// TestLoadConfigSources_ConfigInDirOverlap verifies that a -config file that
+// is also a member of -config-dir is rejected.
+func TestLoadConfigSources_ConfigInDirOverlap(t *testing.T) {
+	dir := t.TempDir()
+	cfgPath := writeYAML(t, dir, "main.yaml", "models:\n"+modelCfg("base", "echo base"))
+
+	_, err := LoadConfigSources(cfgPath, dir)
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "is also present in -config-dir")
+}
+
+func TestLoadConfigSources_DuplicateModelID(t *testing.T) {
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", "models:\n"+modelCfg("dup", "echo a"))
+	writeYAML(t, dir, "b.yaml", "models:\n"+modelCfg("dup", "echo b"))
+
+	_, err := LoadConfigSources("", dir)
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), `duplicate models "dup"`)
+}
+
+func TestLoadConfigSources_DuplicateGroupID(t *testing.T) {
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", `
+models:
+`+modelCfg("m1", "echo m1")+"groups:\n  g1:\n    members: [m1]\n")
+	writeYAML(t, dir, "b.yaml", `
+models:
+`+modelCfg("m2", "echo m2")+"groups:\n  g1:\n    members: [m2]\n")
+
+	_, err := LoadConfigSources("", dir)
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), `duplicate groups "g1"`)
+}
+
+func TestLoadConfigSources_DuplicatePeer(t *testing.T) {
+	dir := t.TempDir()
+	peerA := "peers:\n  remote:\n    proxy: http://x:1\n    models: [m1]\n"
+	peerB := "peers:\n  remote:\n    proxy: http://x:2\n    models: [m2]\n"
+	writeYAML(t, dir, "a.yaml", "models:\n"+modelCfg("m1", "echo m1")+"\n"+peerA)
+	writeYAML(t, dir, "b.yaml", "models:\n"+modelCfg("m2", "echo m2")+"\n"+peerB)
+
+	_, err := LoadConfigSources("", dir)
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), `duplicate peers "remote"`)
+}
+
+func TestLoadConfigSources_ScalarConflict(t *testing.T) {
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", "models:\n"+modelCfg("m1", "echo m1")+"\nglobalTTL: 100\n")
+	writeYAML(t, dir, "b.yaml", "models:\n"+modelCfg("m2", "echo m2")+"\nglobalTTL: 200\n")
+
+	_, err := LoadConfigSources("", dir)
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), `conflict at "globalTTL"`)
+}
+
+func TestLoadConfigSources_ScalarSameValueNoConflict(t *testing.T) {
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", "models:\n"+modelCfg("m1", "echo m1")+"\nglobalTTL: 100\n")
+	writeYAML(t, dir, "b.yaml", "models:\n"+modelCfg("m2", "echo m2")+"\nglobalTTL: 100\n")
+
+	cfg, err := LoadConfigSources("", dir)
+	require.NoError(t, err)
+	assert.Equal(t, 100, cfg.GlobalTTL)
+}
+
+func TestLoadConfigSources_MacrosConcatenate(t *testing.T) {
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", "macros:\n  LOW: 1\nmodels:\n"+modelCfg("m1", "echo ${LOW}"))
+	writeYAML(t, dir, "b.yaml", "macros:\n  HIGH: 2\nmodels:\n"+modelCfg("m2", "echo ${HIGH}"))
+
+	cfg, err := LoadConfigSources("", dir)
+	require.NoError(t, err)
+	// Both macros are available globally after merge.
+	low, ok := cfg.Macros.Get("LOW")
+	require.True(t, ok)
+	assert.Equal(t, 1, low)
+	high, ok := cfg.Macros.Get("HIGH")
+	require.True(t, ok)
+	assert.Equal(t, 2, high)
+}
+
+func TestLoadConfigSources_APIKeysConcatenate(t *testing.T) {
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", "models:\n"+modelCfg("m1", "echo m1")+"\napiKeys: [key-a]\n")
+	writeYAML(t, dir, "b.yaml", "models:\n"+modelCfg("m2", "echo m2")+"\napiKeys: [key-b]\n")
+
+	cfg, err := LoadConfigSources("", dir)
+	require.NoError(t, err)
+	assert.ElementsMatch(t, []string{"key-a", "key-b"}, cfg.RequiredAPIKeys)
+}
+
+func TestLoadConfigSources_RoutingGroupsMerge(t *testing.T) {
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", `
+models:
+`+modelCfg("m1", "echo m1")+`
+routing:
+  router:
+    settings:
+      groups:
+        groupA:
+          members: [m1]
+`)
+	writeYAML(t, dir, "b.yaml", `
+models:
+`+modelCfg("m2", "echo m2")+`
+routing:
+  router:
+    settings:
+      groups:
+        groupB:
+          members: [m2]
+`)
+
+	cfg, err := LoadConfigSources("", dir)
+	require.NoError(t, err)
+	groups := cfg.Routing.Router.Settings.Groups
+	assert.Contains(t, groups, "groupA")
+	assert.Contains(t, groups, "groupB")
+	// default group added by pipeline for orphaned/leftover routing groups...
+	// here both groups reference distinct models
+}
+
+func TestLoadConfigSources_EnvMacrosSubstituted(t *testing.T) {
+	dir := t.TempDir()
+	// Use ${PORT} in cmd so the pipeline allocates a port and substitutes it;
+	// verifies env/macro substitution runs on the merged document.
+	writeYAML(t, dir, "a.yaml", "models:\n  m1:\n    cmd: serve --port ${PORT}\n    proxy: \"http://localhost:${PORT}\"\n")
+	cfg, err := LoadConfigSources("", dir)
+	require.NoError(t, err)
+	m := cfg.Models["m1"]
+	assert.NotContains(t, m.Cmd, "${PORT}", "PORT macro should have been substituted")
+	assert.NotContains(t, m.Proxy, "${PORT}", "PORT macro should have been substituted in proxy")
+}
+
+func TestLoadConfigSources_EnvMacroInFlowStyleList(t *testing.T) {
+	// Regression: flow-style lists with ${env.*} must parse. Previously
+	// parseSource unmarshalled before env substitution, so the brace in
+	// [${env.API_KEY}] was misread as a flow mapping and parsing failed.
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", "models:\n  m1:\n    cmd: echo hi\n    proxy: \"http://localhost:9999\"\n")
+	writeYAML(t, dir, "keys.yaml", "apiKeys: [${env.TEST_API_KEY}]\nmodels:\n  m2:\n    cmd: echo hi\n    proxy: \"http://localhost:9998\"\n")
+
+	t.Setenv("TEST_API_KEY", "secret123")
+	cfg, err := LoadConfigSources("", dir)
+	require.NoError(t, err)
+	assert.Contains(t, cfg.RequiredAPIKeys, "secret123")
+}
+
+func TestLoadConfigSources_SortedOrderDeterministic(t *testing.T) {
+	// Two files defining distinct models, scanned in z..a order by filename.
+	// Determine merged result is the same regardless of how the FS returns them.
+	dir := t.TempDir()
+	writeYAML(t, dir, "z.yaml", "models:\n"+modelCfg("zmodel", "echo z"))
+	writeYAML(t, dir, "a.yaml", "models:\n"+modelCfg("amodel", "echo a"))
+
+	const runs = 3
+	for i := 0; i < runs; i++ {
+		cfg, err := LoadConfigSources("", dir)
+		require.NoError(t, err)
+		// startPort-based allocation: first allocated model gets 5800.
+		// Sorted order means amodel gets 5800, zmodel gets 5801.
+		_, _, ok := cfg.FindConfig("amodel")
+		assert.True(t, ok)
+		_, _, ok = cfg.FindConfig("zmodel")
+		assert.True(t, ok)
+	}
+}
+
+func TestLoadConfigSources_EmptyDirWithConfig(t *testing.T) {
+	dir := t.TempDir()
+	cfgDir := t.TempDir()
+	cfgPath := writeYAML(t, dir, "main.yaml", "models:\n"+modelCfg("m1", "echo m1"))
+
+	cfg, err := LoadConfigSources(cfgPath, cfgDir)
+	require.NoError(t, err)
+	assert.Contains(t, cfg.Models, "m1")
+}
+
+func TestLoadConfigSources_EmptyDirOnly(t *testing.T) {
+	// An empty -config-dir with no -config is an error: there is nothing to
+	// load and silently producing an empty config would mask the misconfig.
+	cfgDir := t.TempDir()
+	_, err := LoadConfigSources("", cfgDir)
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "no configuration sources found")
+}
+
+func TestLoadConfigSources_AssertNoUnknownMacrosAfterMerge(t *testing.T) {
+	// Macros defined in one file should not satisfy unknown-macro validation in
+	// another — they do, because merge concats global macros before validation
+	// runs. This test documents that a macro from file A is usable in file B.
+	dir := t.TempDir()
+	writeYAML(t, dir, "macros.yaml", "macros:\n  SHARED: hello\nmodels:\n"+modelCfg("dummy", "echo dummy"))
+	writeYAML(t, dir, "use.yaml", "models:\n"+modelCfg("user", "echo ${SHARED}"))
+
+	cfg, err := LoadConfigSources("", dir)
+	require.NoError(t, err)
+	m := cfg.Models["user"]
+	assert.Contains(t, m.Cmd, "hello")
+	assert.NotContains(t, m.Cmd, "${SHARED}")
+}
+
+func TestLoadConfigSources_KindMismatchErrors(t *testing.T) {
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", "startPort: 5800\nmodels:\n"+modelCfg("m1", "echo m1"))
+	writeYAML(t, dir, "b.yaml", "startPort: [5800, 5801]\nmodels:\n"+modelCfg("m2", "echo m2"))
+
+	_, err := LoadConfigSources("", dir)
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "incompatible YAML node kinds")
+}
+
+func TestLoadConfigSources_NullYieldsToValue(t *testing.T) {
+	// File A: routing.router block absent (null on root for routing);
+	// file B: defines routing.router.settings.groups. Merge should keep B's.
+	dir := t.TempDir()
+	writeYAML(t, dir, "a.yaml", "models:\n"+modelCfg("m1", "echo m1"))
+	writeYAML(t, dir, "b.yaml", "routing:\n  router:\n    settings:\n      groups:\n        g1:\n          members: [m1]\nmodels:\n"+modelCfg("m2", "echo m2"))
+
+	cfg, err := LoadConfigSources("", dir)
+	require.NoError(t, err)
+	assert.Contains(t, cfg.Routing.Router.Settings.Groups, "g1")
+}

internal/config/model_config.go

@@ -2,6 +2,7 @@ package config
 
 import (
 	"errors"
+	"fmt"
 	"runtime"
 )
 
@@ -9,6 +10,47 @@ const (
 	MODEL_CONFIG_DEFAULT_TTL = -1
 )
 
+var validModalities = map[string]struct{}{
+	"text":  {},
+	"audio": {},
+	"image": {},
+}
+
+// ModelCapConfig defines what modalities and features a model supports.
+// Used in /v1/models to inform clients. An empty block (all zero values) is
+// treated as not configured.
+type ModelCapConfig struct {
+	In       []string `yaml:"in"`
+	Out      []string `yaml:"out"`
+	Tools    bool     `yaml:"tools"`
+	Reranker bool     `yaml:"reranker"`
+	Context  int      `yaml:"context"`
+}
+
+// Empty returns true when all fields are at their zero values.
+func (c ModelCapConfig) Empty() bool {
+	return len(c.In) == 0 && len(c.Out) == 0 && !c.Tools && !c.Reranker && c.Context == 0
+}
+
+// Validate checks that all modality values are recognized and context is
+// non-negative. Returns an error if any value is invalid.
+func (c ModelCapConfig) Validate() error {
+	for _, m := range c.In {
+		if _, ok := validModalities[m]; !ok {
+			return fmt.Errorf("capabilities.in: invalid modality %q, must be one of: text, audio, image", m)
+		}
+	}
+	for _, m := range c.Out {
+		if _, ok := validModalities[m]; !ok {
+			return fmt.Errorf("capabilities.out: invalid modality %q, must be one of: text, audio, image", m)
+		}
+	}
+	if c.Context < 0 {
+		return errors.New("capabilities.context: must be >= 0")
+	}
+	return nil
+}
+
 // TimeoutsConfig holds timeout settings for proxy connections
 // 0 = no timeout
 type TimeoutsConfig struct {
@@ -55,6 +97,9 @@ type ModelConfig struct {
 	// Timeout settings for proxy connections
 	Timeouts TimeoutsConfig `yaml:"timeouts"`
 
+	// Capabilities defines what modalities and features the model supports.
+	Capabilities ModelCapConfig `yaml:"capabilities"`
+
 	// Copy of HealthCheckTimeout from global config
 	HealthCheckTimeout int `yaml:"healthCheckTimeout"`
 }

internal/config/model_config_test.go

@@ -152,7 +152,7 @@ models:
         stop:
           - "<|end|>"
           - "<|stop|>"
-`
+ `
 	config, err := LoadConfigFromReader(strings.NewReader(content))
 	assert.NoError(t, err)
 
@@ -170,3 +170,167 @@ models:
 	assert.Equal(t, 0.7, setParams["temperature"])
 	assert.Equal(t, 0.9, setParams["top_p"])
 }
+
+func TestConfig_ModelCapabilities(t *testing.T) {
+	t.Run("all fields", func(t *testing.T) {
+		content := `
+models:
+  model1:
+    cmd: path/to/cmd --port ${PORT}
+    capabilities:
+      in:
+        - text
+        - audio
+        - image
+      out:
+        - text
+        - audio
+        - image
+      tools: true
+      context: 32000
+`
+		config, err := LoadConfigFromReader(strings.NewReader(content))
+		assert.NoError(t, err)
+
+		mc := config.Models["model1"]
+		assert.False(t, mc.Capabilities.Empty())
+		assert.Equal(t, []string{"text", "audio", "image"}, mc.Capabilities.In)
+		assert.Equal(t, []string{"text", "audio", "image"}, mc.Capabilities.Out)
+		assert.True(t, mc.Capabilities.Tools)
+		assert.Equal(t, 32000, mc.Capabilities.Context)
+	})
+
+	t.Run("partial fields", func(t *testing.T) {
+		content := `
+models:
+  model1:
+    cmd: path/to/cmd --port ${PORT}
+    capabilities:
+      tools: true
+      context: 8192
+`
+		config, err := LoadConfigFromReader(strings.NewReader(content))
+		assert.NoError(t, err)
+
+		mc := config.Models["model1"]
+		assert.False(t, mc.Capabilities.Empty())
+		assert.Nil(t, mc.Capabilities.In)
+		assert.Nil(t, mc.Capabilities.Out)
+		assert.True(t, mc.Capabilities.Tools)
+		assert.Equal(t, 8192, mc.Capabilities.Context)
+	})
+
+	t.Run("not set", func(t *testing.T) {
+		content := `
+models:
+  model1:
+    cmd: path/to/cmd --port ${PORT}
+`
+		config, err := LoadConfigFromReader(strings.NewReader(content))
+		assert.NoError(t, err)
+
+		mc := config.Models["model1"]
+		assert.True(t, mc.Capabilities.Empty())
+	})
+
+	t.Run("tools false is empty", func(t *testing.T) {
+		content := `
+models:
+  model1:
+    cmd: path/to/cmd --port ${PORT}
+    capabilities:
+      tools: false
+`
+		config, err := LoadConfigFromReader(strings.NewReader(content))
+		assert.NoError(t, err)
+
+		mc := config.Models["model1"]
+		assert.True(t, mc.Capabilities.Empty())
+	})
+
+	t.Run("reranker true is not empty", func(t *testing.T) {
+		content := `
+models:
+  model1:
+    cmd: path/to/cmd --port ${PORT}
+    capabilities:
+      reranker: true
+`
+		config, err := LoadConfigFromReader(strings.NewReader(content))
+		assert.NoError(t, err)
+
+		mc := config.Models["model1"]
+		assert.False(t, mc.Capabilities.Empty())
+		assert.True(t, mc.Capabilities.Reranker)
+	})
+
+	t.Run("reranker false is empty", func(t *testing.T) {
+		content := `
+models:
+  model1:
+    cmd: path/to/cmd --port ${PORT}
+    capabilities:
+      reranker: false
+`
+		config, err := LoadConfigFromReader(strings.NewReader(content))
+		assert.NoError(t, err)
+
+		mc := config.Models["model1"]
+		assert.True(t, mc.Capabilities.Empty())
+	})
+}
+
+func TestConfig_ModelCapabilities_Validate(t *testing.T) {
+	t.Run("valid_modalities", func(t *testing.T) {
+		caps := ModelCapConfig{
+			In:      []string{"text", "image"},
+			Out:     []string{"text", "audio"},
+			Tools:   true,
+			Context: 100000,
+		}
+		assert.NoError(t, caps.Validate())
+	})
+
+	t.Run("empty_is_valid", func(t *testing.T) {
+		caps := ModelCapConfig{}
+		assert.NoError(t, caps.Validate())
+	})
+
+	t.Run("invalid_in_modality", func(t *testing.T) {
+		caps := ModelCapConfig{In: []string{"video"}}
+		err := caps.Validate()
+		assert.Error(t, err)
+		assert.Contains(t, err.Error(), "capabilities.in")
+		assert.Contains(t, err.Error(), "video")
+	})
+
+	t.Run("invalid_out_modality", func(t *testing.T) {
+		caps := ModelCapConfig{Out: []string{"video"}}
+		err := caps.Validate()
+		assert.Error(t, err)
+		assert.Contains(t, err.Error(), "capabilities.out")
+		assert.Contains(t, err.Error(), "video")
+	})
+
+	t.Run("negative_context", func(t *testing.T) {
+		caps := ModelCapConfig{Context: -1}
+		err := caps.Validate()
+		assert.Error(t, err)
+		assert.Contains(t, err.Error(), "capabilities.context")
+	})
+
+	t.Run("rejects_invalid_at_load", func(t *testing.T) {
+		content := `
+models:
+  model1:
+    cmd: path/to/cmd --port ${PORT}
+    capabilities:
+      in:
+        - text
+        - video
+`
+		_, err := LoadConfigFromReader(strings.NewReader(content))
+		assert.Error(t, err)
+		assert.Contains(t, err.Error(), "video")
+	})
+}

internal/config/upstream.go

@@ -0,0 +1,55 @@
+package config
+
+import (
+	"fmt"
+	"regexp"
+
+	"gopkg.in/yaml.v3"
+)
+
+// DefaultUpstreamIgnorePathsPattern is the default regular expression applied
+// to upstream.ignorePaths when the section is empty or absent from the config.
+// It matches common static-asset suffixes so requests for .js/.css/.png/etc.
+// files do not trigger a model swap.
+const DefaultUpstreamIgnorePathsPattern = `.*\.(js|json|css|png|gif|jpg|jpeg|ico|txt)$`
+
+// DefaultUpstreamIgnorePaths returns the default compiled ignore paths used
+// when upstream.ignorePaths is not specified in the config. The returned slice
+// is fresh so callers may mutate it without affecting other configs.
+func DefaultUpstreamIgnorePaths() []*regexp.Regexp {
+	return []*regexp.Regexp{regexp.MustCompile(DefaultUpstreamIgnorePathsPattern)}
+}
+
+// UpstreamConfig controls behaviour of the /upstream passthrough endpoint.
+type UpstreamConfig struct {
+	// IgnorePaths is a slice of compiled regular expressions. Any request to
+	// /upstream/<model>/<path> whose remaining path matches any of these
+	// expressions will be ignored and not trigger a swap. When the config
+	// does not specify any patterns, DefaultUpstreamIgnorePaths is applied.
+	IgnorePaths []*regexp.Regexp `yaml:"-"`
+}
+
+// rawUpstreamConfig is the intermediate form used to unmarshal the YAML into
+// plain strings, which are then compiled into *regexp.Regexp.
+type rawUpstreamConfig struct {
+	IgnorePaths []string `yaml:"ignorePaths"`
+}
+
+// UnmarshalYAML compiles each ignorePaths entry into a *regexp.Regexp. If any
+// entry fails to compile, an error is returned.
+func (u *UpstreamConfig) UnmarshalYAML(value *yaml.Node) error {
+	var raw rawUpstreamConfig
+	if err := value.Decode(&raw); err != nil {
+		return err
+	}
+	patterns := make([]*regexp.Regexp, 0, len(raw.IgnorePaths))
+	for _, p := range raw.IgnorePaths {
+		re, err := regexp.Compile(p)
+		if err != nil {
+			return fmt.Errorf("upstream.ignorePaths: invalid regular expression %q: %w", p, err)
+		}
+		patterns = append(patterns, re)
+	}
+	u.IgnorePaths = patterns
+	return nil
+}

internal/config/upstream_test.go

@@ -0,0 +1,88 @@
+package config
+
+import (
+	"regexp"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+const upstreamConfigHeader = `
+models:
+  model1:
+    cmd: path/to/cmd --arg1 one
+    proxy: "http://localhost:8080"
+`
+
+func TestConfig_UpstreamIgnorePaths_DefaultWhenAbsent(t *testing.T) {
+	// When upstream is not specified at all, the default pattern is applied.
+	content := upstreamConfigHeader
+	cfg, err := LoadConfigFromReader(strings.NewReader(content))
+	require.NoError(t, err)
+	require.Len(t, cfg.Upstream.IgnorePaths, 1)
+
+	def := cfg.Upstream.IgnorePaths[0]
+	assert.IsType(t, &regexp.Regexp{}, def)
+	assert.Equal(t, DefaultUpstreamIgnorePathsPattern, def.String())
+
+	// The default matches common static-asset suffixes.
+	assert.True(t, def.MatchString("/foo.js"))
+	assert.True(t, def.MatchString("/bar/baz.json"))
+	assert.True(t, def.MatchString("/static/img.png"))
+	assert.True(t, def.MatchString("/notes.txt"))
+	assert.True(t, def.MatchString("/favicon.ico"))
+	// And does not match inference API paths.
+	assert.False(t, def.MatchString("/v1/chat/completions"))
+	assert.False(t, def.MatchString("/v1/models"))
+	assert.False(t, def.MatchString("/health"))
+}
+
+func TestConfig_UpstreamIgnorePaths_DefaultWhenSectionEmpty(t *testing.T) {
+	// When upstream is present but ignorePaths is omitted, the default is still
+	// applied.
+	content := `upstream: {}` + "\n" + upstreamConfigHeader
+	cfg, err := LoadConfigFromReader(strings.NewReader(content))
+	require.NoError(t, err)
+	require.Len(t, cfg.Upstream.IgnorePaths, 1)
+	assert.Equal(t, DefaultUpstreamIgnorePathsPattern, cfg.Upstream.IgnorePaths[0].String())
+}
+
+func TestConfig_UpstreamIgnorePaths_Compiles(t *testing.T) {
+	content := `
+upstream:
+  ignorePaths:
+    - ".*\\.(js|json|css|png|gif|jpg|jpeg|txt)$"
+    - "^/static/.*"
+` + upstreamConfigHeader
+
+	cfg, err := LoadConfigFromReader(strings.NewReader(content))
+	require.NoError(t, err)
+	require.Len(t, cfg.Upstream.IgnorePaths, 2)
+
+	// Verify the patterns are compiled into *regexp.Regexp and match as expected.
+	assert.True(t, cfg.Upstream.IgnorePaths[0].MatchString("/foo.js"))
+	assert.True(t, cfg.Upstream.IgnorePaths[0].MatchString("/bar/baz.json"))
+	assert.False(t, cfg.Upstream.IgnorePaths[0].MatchString("/v1/chat/completions"))
+	assert.True(t, cfg.Upstream.IgnorePaths[1].MatchString("/static/foo.png"))
+	assert.False(t, cfg.Upstream.IgnorePaths[1].MatchString("/v1/chat/completions"))
+
+	// Confirm the type is *regexp.Regexp to satisfy the API contract.
+	for _, re := range cfg.Upstream.IgnorePaths {
+		assert.IsType(t, &regexp.Regexp{}, re)
+	}
+}
+
+func TestConfig_UpstreamIgnorePaths_InvalidRegexReturnsError(t *testing.T) {
+	content := `
+upstream:
+  ignorePaths:
+    - "[invalid("
+` + upstreamConfigHeader
+
+	_, err := LoadConfigFromReader(strings.NewReader(content))
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "upstream.ignorePaths")
+	assert.Contains(t, err.Error(), "invalid regular expression")
+}

internal/perf/d3dkmt_types.go

@@ -0,0 +1,92 @@
+package perf
+
+type LUID struct {
+	LowPart  uint32
+	HighPart int32
+}
+
+const maxEnumAdapters = 16
+
+type D3DKMT_ENUMADAPTERS2 struct {
+	NumAdapters uint32
+	pAdapters   uintptr
+}
+
+type D3DKMT_ADAPTERINFO struct {
+	hAdapter                     uint32
+	AdapterLuid                  LUID
+	NumOfSources                 uint32
+	bPresentMoveRegionsPreferred int32
+}
+
+type D3DKMT_OPENADAPTERFROMLUID struct {
+	AdapterLuid LUID
+	hAdapter    uint32
+}
+
+type D3DKMT_CLOSEADAPTER struct {
+	hAdapter uint32
+}
+
+type KMTQUERYADAPTERINFOTYPE int32
+
+const (
+	KMTQAITYPE_UMDRIVERPRIVATE          KMTQUERYADAPTERINFOTYPE = 0
+	KMTQAITYPE_ADAPTERREGISTRYINFO      KMTQUERYADAPTERINFOTYPE = 8
+	KMTQAITYPE_DRIVERVERSION            KMTQUERYADAPTERINFOTYPE = 13
+	KMTQAITYPE_PHYSICALADAPTERDEVICEIDS KMTQUERYADAPTERINFOTYPE = 31
+	KMTQAITYPE_NODEPERFDATA             KMTQUERYADAPTERINFOTYPE = 61
+	KMTQAITYPE_ADAPTERPERFDATA          KMTQUERYADAPTERINFOTYPE = 62
+	KMTQAITYPE_ADAPTERPERFDATA_CAPS     KMTQUERYADAPTERINFOTYPE = 63
+)
+
+type D3DKMT_QUERYADAPTERINFO struct {
+	hAdapter              uint32
+	Type                  KMTQUERYADAPTERINFOTYPE
+	pPrivateDriverData    uintptr
+	PrivateDriverDataSize uint32
+}
+
+type D3DKMT_ADAPTER_PERFDATA struct {
+	PhysicalAdapterIndex uint32
+	MemoryFrequency      uint64
+	MaxMemoryFrequency   uint64
+	MaxMemoryFrequencyOC uint64
+	MemoryBandwidth      uint64
+	PCIEBandwidth        uint64
+	FanRPM               uint32
+	Power                uint32
+	Temperature          uint32
+	PowerStateOverride   byte
+}
+
+type D3DKMT_QUERYSTATISTICS_TYPE int32
+
+const (
+	D3DKMT_QUERYSTATISTICS_ADAPTER             D3DKMT_QUERYSTATISTICS_TYPE = 0
+	D3DKMT_QUERYSTATISTICS_PROCESS             D3DKMT_QUERYSTATISTICS_TYPE = 1
+	D3DKMT_QUERYSTATISTICS_PROCESS_ADAPTER     D3DKMT_QUERYSTATISTICS_TYPE = 2
+	D3DKMT_QUERYSTATISTICS_SEGMENT             D3DKMT_QUERYSTATISTICS_TYPE = 3
+	D3DKMT_QUERYSTATISTICS_PROCESS_SEGMENT     D3DKMT_QUERYSTATISTICS_TYPE = 4
+	D3DKMT_QUERYSTATISTICS_NODE                D3DKMT_QUERYSTATISTICS_TYPE = 5
+	D3DKMT_QUERYSTATISTICS_PROCESS_NODE        D3DKMT_QUERYSTATISTICS_TYPE = 6
+	D3DKMT_QUERYSTATISTICS_VIDPNSOURCE         D3DKMT_QUERYSTATISTICS_TYPE = 7
+	D3DKMT_QUERYSTATISTICS_PROCESS_VIDPNSOURCE D3DKMT_QUERYSTATISTICS_TYPE = 8
+)
+
+type D3DKMT_ADAPTER_PERFDATACAPS struct {
+	PhysicalAdapterIndex uint32
+	MaxMemoryBandwidth   uint64
+	MaxPCIEBandwidth     uint64
+	MaxFanRPM            uint32
+	TemperatureMax       uint32
+	TemperatureWarning   uint32
+}
+
+type D3DKMT_QUERYSTATISTICS_QUERY_SEGMENT struct {
+	SegmentId uint32
+}
+
+type D3DKMT_QUERYSTATISTICS_QUERY_NODE struct {
+	NodeId uint32
+}

internal/perf/d3dkmt_windows.go

@@ -0,0 +1,529 @@
+//go:build windows
+
+package perf
+
+import (
+	"context"
+	"encoding/binary"
+	"fmt"
+	"sync"
+	"time"
+	"unsafe"
+
+	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"golang.org/x/sys/windows"
+)
+
+var (
+	d3dkmDLL                *windows.LazyDLL
+	procEnumAdapters2       *windows.LazyProc
+	procOpenAdapterFromLuid *windows.LazyProc
+	procCloseAdapter        *windows.LazyProc
+	procQueryAdapterInfo    *windows.LazyProc
+	procQueryStatistics     *windows.LazyProc
+	d3dkmtInitOnce          sync.Once
+	d3dkmtInitErr           error
+)
+
+// initD3DKMT lazily loads gdi32.dll and resolves D3DKMT function pointers.
+// Safe for concurrent use via sync.Once.
+func initD3DKMT() error {
+	d3dkmtInitOnce.Do(func() {
+		d3dkmDLL = windows.NewLazySystemDLL("gdi32.dll")
+
+		procEnumAdapters2 = d3dkmDLL.NewProc("D3DKMTEnumAdapters2")
+		procOpenAdapterFromLuid = d3dkmDLL.NewProc("D3DKMTOpenAdapterFromLuid")
+		procCloseAdapter = d3dkmDLL.NewProc("D3DKMTCloseAdapter")
+		procQueryAdapterInfo = d3dkmDLL.NewProc("D3DKMTQueryAdapterInfo")
+		procQueryStatistics = d3dkmDLL.NewProc("D3DKMTQueryStatistics")
+
+		for name, p := range map[string]*windows.LazyProc{
+			"D3DKMTEnumAdapters2":       procEnumAdapters2,
+			"D3DKMTOpenAdapterFromLuid": procOpenAdapterFromLuid,
+			"D3DKMTCloseAdapter":        procCloseAdapter,
+			"D3DKMTQueryAdapterInfo":    procQueryAdapterInfo,
+			"D3DKMTQueryStatistics":     procQueryStatistics,
+		} {
+			if err := p.Find(); err != nil {
+				d3dkmtInitErr = fmt.Errorf("D3DKMT %s not found: %w", name, err)
+				return
+			}
+		}
+	})
+	return d3dkmtInitErr
+}
+
+// ntstatusCall invokes a D3DKMT function and returns a non-nil error if the
+// NTSTATUS result is not STATUS_SUCCESS (0).
+func ntstatusCall(proc *windows.LazyProc, arg unsafe.Pointer) error {
+	ret, _, _ := proc.Call(uintptr(arg))
+	if ret != 0 {
+		return fmt.Errorf("NTSTATUS 0x%08x", uint32(ret))
+	}
+	return nil
+}
+
+// d3dkmEnumerateAdapters enumerates all available graphics adapters via
+// D3DKMTEnumAdapters2.
+func d3dkmEnumerateAdapters() ([]D3DKMT_ADAPTERINFO, error) {
+	var adapters [maxEnumAdapters]D3DKMT_ADAPTERINFO
+	enum := D3DKMT_ENUMADAPTERS2{
+		NumAdapters: maxEnumAdapters,
+		pAdapters:   uintptr(unsafe.Pointer(&adapters[0])),
+	}
+	if err := ntstatusCall(procEnumAdapters2, unsafe.Pointer(&enum)); err != nil {
+		return nil, fmt.Errorf("EnumAdapters2: %w", err)
+	}
+	if enum.NumAdapters == 0 {
+		return nil, fmt.Errorf("no adapters found")
+	}
+	result := make([]D3DKMT_ADAPTERINFO, enum.NumAdapters)
+	for i := uint32(0); i < enum.NumAdapters; i++ {
+		result[i] = adapters[i]
+	}
+	return result, nil
+}
+
+// d3dkmOpenAdapter opens a D3DKMT adapter handle for the given LUID.
+func d3dkmOpenAdapter(luid LUID) (uint32, error) {
+	req := D3DKMT_OPENADAPTERFROMLUID{
+		AdapterLuid: luid,
+	}
+	if err := ntstatusCall(procOpenAdapterFromLuid, unsafe.Pointer(&req)); err != nil {
+		return 0, fmt.Errorf("OpenAdapterFromLuid: %w", err)
+	}
+	return req.hAdapter, nil
+}
+
+// d3dkmCloseAdapter closes a previously opened D3DKMT adapter handle.
+func d3dkmCloseAdapter(hAdapter uint32) error {
+	req := D3DKMT_CLOSEADAPTER{hAdapter: hAdapter}
+	return ntstatusCall(procCloseAdapter, unsafe.Pointer(&req))
+}
+
+// d3dkmGetAdapterPerfData queries per-adapter performance data (temperature,
+// fan RPM, power, bandwidth) via KMTQAITYPE_ADAPTERPERFDATA.
+func d3dkmGetAdapterPerfData(hAdapter uint32) (*D3DKMT_ADAPTER_PERFDATA, error) {
+	var data D3DKMT_ADAPTER_PERFDATA
+	req := D3DKMT_QUERYADAPTERINFO{
+		hAdapter:              hAdapter,
+		Type:                  KMTQAITYPE_ADAPTERPERFDATA,
+		pPrivateDriverData:    uintptr(unsafe.Pointer(&data)),
+		PrivateDriverDataSize: uint32(unsafe.Sizeof(data)),
+	}
+	if err := ntstatusCall(procQueryAdapterInfo, unsafe.Pointer(&req)); err != nil {
+		return nil, fmt.Errorf("QueryAdapterInfo(ADAPTERPERFDATA): %w", err)
+	}
+	return &data, nil
+}
+
+// d3dkmGetAdapterPerfDataCaps queries static adapter performance capabilities
+// (max fan RPM, temperature limits, max bandwidth) via KMTQAITYPE_ADAPTERPERFDATA_CAPS.
+func d3dkmGetAdapterPerfDataCaps(hAdapter uint32) (*D3DKMT_ADAPTER_PERFDATACAPS, error) {
+	var data D3DKMT_ADAPTER_PERFDATACAPS
+	req := D3DKMT_QUERYADAPTERINFO{
+		hAdapter:              hAdapter,
+		Type:                  KMTQAITYPE_ADAPTERPERFDATA_CAPS,
+		pPrivateDriverData:    uintptr(unsafe.Pointer(&data)),
+		PrivateDriverDataSize: uint32(unsafe.Sizeof(data)),
+	}
+	if err := ntstatusCall(procQueryAdapterInfo, unsafe.Pointer(&req)); err != nil {
+		return nil, fmt.Errorf("QueryAdapterInfo(ADAPTERPERFDATACAPS): %w", err)
+	}
+	return &data, nil
+}
+
+type queryStatsBuffer struct {
+	Type        int32   // offset 0
+	AdapterLuid LUID    // offset 4
+	hProcess    uintptr // offset 16
+	// _result mirrors the D3DKMT_QUERYSTATISTICS_RESULT union.
+	// sizeof(D3DKMT_QUERYSTATISTICS) == 0x328 (808 bytes) on x64.
+	//
+	// The C struct layout (x64):
+	//   offset  0: Type (int32, 4 bytes)
+	//   offset  4: AdapterLuid (LUID, 8 bytes)
+	//   offset 12: 4 bytes padding (for 8-byte alignment of hProcess)
+	//   offset 16: hProcess (HANDLE, 8 bytes)
+	//   offset 24: QueryResult (union, 780 bytes — largest member is AdapterInformation)
+	//   offset 804: anonymous input union (QueryNode.NodeId / QuerySegment.SegmentId, 4 bytes)
+	//
+	// Previous bug: _result was [776]byte, placing QueryId at offset 800 instead of 804.
+	// The kernel read NodeId/SegmentId from offset 804 (always zero from _pad),
+	// causing all NODE and SEGMENT queries to use index 0 regardless of the value
+	// passed in QueryId. This produced alternating behavior where only GPU util OR
+	// memory util appeared to work, depending on which test variant happened to put
+	// non-zero data near offset 804 in the result buffer.
+	_result [780]byte // offset 24, size 780 — places QueryId at offset 804
+	QueryId int32     // offset 804 — matches C anonymous union for NodeId/SegmentId
+}
+
+func init() {
+	var buf queryStatsBuffer
+	if unsafe.Sizeof(buf) != 808 {
+		panic(fmt.Sprintf("queryStatsBuffer size %d != expected 808 (sizeof D3DKMT_QUERYSTATISTICS on x64)", unsafe.Sizeof(buf)))
+	}
+	if unsafe.Offsetof(buf.QueryId) != 804 {
+		panic(fmt.Sprintf("queryStatsBuffer.QueryId offset %d != expected 804 (C anonymous union offset)", unsafe.Offsetof(buf.QueryId)))
+	}
+
+	var perfData D3DKMT_ADAPTER_PERFDATA
+	if unsafe.Sizeof(perfData) != 64 {
+		panic(fmt.Sprintf("D3DKMT_ADAPTER_PERFDATA size %d != expected 64 on x64", unsafe.Sizeof(perfData)))
+	}
+
+	var caps D3DKMT_ADAPTER_PERFDATACAPS
+	if unsafe.Sizeof(caps) != 40 {
+		panic(fmt.Sprintf("D3DKMT_ADAPTER_PERFDATACAPS size %d != expected 40 on x64", unsafe.Sizeof(caps)))
+	}
+}
+
+const (
+	qsoffsetNbSegments        = 0
+	qsoffsetNodeCount         = 4
+	qsoffsetCommitLimit       = 0
+	qsoffsetBytesCommitted    = 8
+	qsoffsetBytesResident     = 16
+	qsoffsetRunningTime       = 0
+	qsoffsetSystemRunningTime = 272
+)
+
+// d3dkmQueryAdapterStats returns the number of memory segments and compute
+// nodes for the adapter identified by luid.
+func d3dkmQueryAdapterStats(luid LUID) (nbSegments uint32, nodeCount uint32, err error) {
+	buf := queryStatsBuffer{
+		Type:        int32(D3DKMT_QUERYSTATISTICS_ADAPTER),
+		AdapterLuid: luid,
+	}
+	if err := ntstatusCall(procQueryStatistics, unsafe.Pointer(&buf)); err != nil {
+		return 0, 0, fmt.Errorf("QueryStatistics(ADAPTER): %w", err)
+	}
+	nbSegments = binary.LittleEndian.Uint32(buf._result[qsoffsetNbSegments : qsoffsetNbSegments+4])
+	nodeCount = binary.LittleEndian.Uint32(buf._result[qsoffsetNodeCount : qsoffsetNodeCount+4])
+	return nbSegments, nodeCount, nil
+}
+
+// d3dkmQuerySegmentStats returns the commit limit (total) and resident
+// (used) bytes for the given memory segment of an adapter.
+func d3dkmQuerySegmentStats(luid LUID, segmentID uint32) (commitLimit uint64, bytesResident uint64, err error) {
+	buf := queryStatsBuffer{
+		Type:        int32(D3DKMT_QUERYSTATISTICS_SEGMENT),
+		AdapterLuid: luid,
+		QueryId:     int32(segmentID),
+	}
+	if err := ntstatusCall(procQueryStatistics, unsafe.Pointer(&buf)); err != nil {
+		return 0, 0, fmt.Errorf("QueryStatistics(SEGMENT %d): %w", segmentID, err)
+	}
+	commitLimit = binary.LittleEndian.Uint64(buf._result[qsoffsetCommitLimit : qsoffsetCommitLimit+8])
+	bytesResident = binary.LittleEndian.Uint64(buf._result[qsoffsetBytesResident : qsoffsetBytesResident+8])
+	if bytesResident == 0 {
+		bytesResident = binary.LittleEndian.Uint64(buf._result[qsoffsetBytesCommitted : qsoffsetBytesCommitted+8])
+	}
+	return commitLimit, bytesResident, nil
+}
+
+// d3dkmQueryNodeStats returns the global and system running time counters
+// (in 100ns units) for the given compute node of an adapter.
+func d3dkmQueryNodeStats(luid LUID, nodeID uint32) (runningTime uint64, systemRunningTime uint64, err error) {
+	buf := queryStatsBuffer{
+		Type:        int32(D3DKMT_QUERYSTATISTICS_NODE),
+		AdapterLuid: luid,
+		QueryId:     int32(nodeID),
+	}
+	if err := ntstatusCall(procQueryStatistics, unsafe.Pointer(&buf)); err != nil {
+		return 0, 0, fmt.Errorf("QueryStatistics(NODE %d): %w", nodeID, err)
+	}
+	runningTime = binary.LittleEndian.Uint64(buf._result[qsoffsetRunningTime : qsoffsetRunningTime+8])
+	systemRunningTime = binary.LittleEndian.Uint64(buf._result[qsoffsetSystemRunningTime : qsoffsetSystemRunningTime+8])
+	return runningTime, systemRunningTime, nil
+}
+
+type nodeRunningTimes struct {
+	Global uint64
+	System uint64
+}
+
+// d3dkmtNodeUtil computes GPU node utilization as a percentage from running
+// time deltas. Returns -1 if counters went backwards (wrap/reset), 0 if idle.
+func d3dkmtNodeUtil(prevRT, curRT nodeRunningTimes, elapsed100ns int64) float64 {
+	if curRT.Global < prevRT.Global || curRT.System < prevRT.System {
+		return -1
+	}
+	gd := curRT.Global - prevRT.Global
+	sd := curRT.System - prevRT.System
+
+	if gd > 0 && sd > 0 {
+		util := float64(gd) / float64(sd)
+		if util > 1.0 {
+			util = 1.0
+		}
+		return util * 100.0
+	} else if gd > 0 && elapsed100ns > 0 {
+		util := float64(gd) / float64(elapsed100ns) * 100.0
+		if util > 100.0 {
+			util = 100.0
+		}
+		return util
+	}
+	return 0
+}
+
+// d3dkmtFanPct returns fan speed as a percentage of maxFanRPM, clamped to
+// 100%. Returns 0 if maxFanRPM is unavailable or fan is not spinning.
+func d3dkmtFanPct(fanRPM, maxFanRPM uint32) float64 {
+	if maxFanRPM > 0 && fanRPM > 0 {
+		pct := float64(fanRPM) / float64(maxFanRPM) * 100.0
+		if pct > 100.0 {
+			pct = 100.0
+		}
+		return pct
+	}
+	return 0
+}
+
+// d3dkmtPowerW converts power from deci-watts (as reported by D3DKMT) to
+// watts. Returns 0 if the power value is zero.
+func d3dkmtPowerW(power uint32) float64 {
+	if power > 0 {
+		return float64(power) / 10.0
+	}
+	return 0
+}
+
+// d3dkmtTempC converts temperature from deci-Celsius (as reported by D3DKMT)
+// to degrees Celsius.
+func d3dkmtTempC(tempDeciC uint32) int {
+	return int(tempDeciC / 10)
+}
+
+type d3dkmtAdapterState struct {
+	luid       LUID
+	hAdapter   uint32
+	nbSegments uint32
+	nodeCount  uint32
+	maxFanRPM  uint32
+	prevNodeRT map[uint32]nodeRunningTimes
+	prevTime   time.Time
+}
+
+// tryD3DKMT attempts to start GPU monitoring using D3DKMT and optional PDH
+// counters. It returns a channel of GpuStat snapshots or an error if no
+// usable adapters are found.
+func tryD3DKMT(ctx context.Context, every time.Duration, logger *logmon.Monitor) (chan []GpuStat, error) {
+	if err := initD3DKMT(); err != nil {
+		return nil, err
+	}
+
+	adapterInfos, err := d3dkmEnumerateAdapters()
+	if err != nil {
+		return nil, err
+	}
+
+	type adapterMeta struct {
+		luid       LUID
+		nbSegments uint32
+		nodeCount  uint32
+		maxFanRPM  uint32
+	}
+
+	var metaList []adapterMeta
+
+	for i, ai := range adapterInfos {
+		hAdapter, err := d3dkmOpenAdapter(ai.AdapterLuid)
+		if err != nil {
+			logger.Debugf("adapter %d: open failed: %s", i, err.Error())
+			continue
+		}
+
+		nbSegments, nodeCount, err := d3dkmQueryAdapterStats(ai.AdapterLuid)
+		if err != nil {
+			logger.Debugf("adapter %d: query stats failed: %s", i, err.Error())
+			d3dkmCloseAdapter(hAdapter)
+			continue
+		}
+
+		caps, err := d3dkmGetAdapterPerfDataCaps(hAdapter)
+		if err != nil {
+			logger.Debugf("adapter %d: perf caps failed: %s", i, err.Error())
+		}
+
+		d3dkmCloseAdapter(hAdapter)
+
+		var maxFanRPM uint32
+		if caps != nil {
+			maxFanRPM = caps.MaxFanRPM
+		}
+
+		metaList = append(metaList, adapterMeta{
+			luid:       ai.AdapterLuid,
+			nbSegments: nbSegments,
+			nodeCount:  nodeCount,
+			maxFanRPM:  maxFanRPM,
+		})
+		logger.Debugf("adapter %d: segments=%d nodes=%d fan_max=%d luid=%d:%d", i, nbSegments, nodeCount, maxFanRPM, ai.AdapterLuid.HighPart, ai.AdapterLuid.LowPart)
+	}
+
+	if len(metaList) == 0 {
+		return nil, fmt.Errorf("no usable D3DKMT adapters found")
+	}
+
+	pdhUtil, pdhErr := initPdhGpuUtil()
+	if pdhErr != nil {
+		logger.Debugf("PDH GPU utilization not available: %s", pdhErr.Error())
+	} else {
+		logger.Info("using PDH performance counters for GPU utilization")
+	}
+
+	ch := make(chan []GpuStat, 1)
+
+	go func() {
+		defer close(ch)
+		if pdhUtil != nil {
+			defer pdhUtil.close()
+		}
+
+		var adapters []d3dkmtAdapterState
+		for _, m := range metaList {
+			hAdapter, err := d3dkmOpenAdapter(m.luid)
+			if err != nil {
+				logger.Debugf("reopen adapter failed: %s", err.Error())
+				continue
+			}
+			adapters = append(adapters, d3dkmtAdapterState{
+				luid:       m.luid,
+				hAdapter:   hAdapter,
+				nbSegments: m.nbSegments,
+				nodeCount:  m.nodeCount,
+				maxFanRPM:  m.maxFanRPM,
+				prevNodeRT: make(map[uint32]nodeRunningTimes),
+			})
+		}
+
+		if len(adapters) == 0 {
+			return
+		}
+
+		defer func() {
+			for _, a := range adapters {
+				d3dkmCloseAdapter(a.hAdapter)
+			}
+		}()
+
+		for i := range adapters {
+			a := &adapters[i]
+			for node := uint32(0); node < a.nodeCount; node++ {
+				globalRT, systemRT, err := d3dkmQueryNodeStats(a.luid, node)
+				if err != nil {
+					continue
+				}
+				a.prevNodeRT[node] = nodeRunningTimes{Global: globalRT, System: systemRT}
+			}
+			a.prevTime = time.Now()
+		}
+
+		ticker := time.NewTicker(every)
+		defer ticker.Stop()
+
+		for {
+			select {
+			case <-ctx.Done():
+				return
+			case <-ticker.C:
+				stats := make([]GpuStat, 0, len(adapters))
+				now := time.Now()
+
+				var pdhUtilMap map[LUID]float64
+				if pdhUtil != nil {
+					pdhUtilMap = pdhUtil.collect()
+				}
+
+				for i := range adapters {
+					a := &adapters[i]
+
+					perfData, err := d3dkmGetAdapterPerfData(a.hAdapter)
+					if err != nil {
+						logger.Debugf("adapter %d perfdata: %s", i, err.Error())
+						continue
+					}
+
+					var memUsedMB, memTotalMB int
+					for seg := uint32(0); seg < a.nbSegments; seg++ {
+						limit, resident, err := d3dkmQuerySegmentStats(a.luid, seg)
+						if err != nil {
+							continue
+						}
+						memUsedMB += int(resident / (1024 * 1024))
+						memTotalMB += int(limit / (1024 * 1024))
+					}
+
+					var gpuUtil float64
+					pdhGaveValue := false
+					if pdhUtilMap != nil {
+						if util, ok := pdhUtilMap[a.luid]; ok {
+							gpuUtil = util
+							pdhGaveValue = true
+						}
+					}
+
+					if !pdhGaveValue && a.nodeCount > 0 {
+						elapsedNs := now.Sub(a.prevTime).Nanoseconds()
+						elapsed100ns := elapsedNs / 100
+
+						for node := uint32(0); node < a.nodeCount; node++ {
+							globalRT, systemRT, err := d3dkmQueryNodeStats(a.luid, node)
+							if err != nil {
+								continue
+							}
+
+							if prevRT, ok := a.prevNodeRT[node]; ok {
+								if globalRT < prevRT.Global || systemRT < prevRT.System {
+									a.prevNodeRT[node] = nodeRunningTimes{Global: globalRT, System: systemRT}
+									continue
+								}
+								nodeUtil := d3dkmtNodeUtil(prevRT, nodeRunningTimes{Global: globalRT, System: systemRT}, elapsed100ns)
+								if nodeUtil > gpuUtil {
+									gpuUtil = nodeUtil
+								}
+							}
+							a.prevNodeRT[node] = nodeRunningTimes{Global: globalRT, System: systemRT}
+						}
+
+						a.prevTime = now
+					}
+
+					tempC := d3dkmtTempC(perfData.Temperature)
+
+					fanSpeedPct := d3dkmtFanPct(perfData.FanRPM, a.maxFanRPM)
+					powerDrawW := d3dkmtPowerW(perfData.Power)
+
+					var memUtilPct float64
+					if memTotalMB > 0 {
+						memUtilPct = float64(memUsedMB) / float64(memTotalMB) * 100.0
+					}
+
+					stats = append(stats, GpuStat{
+						Timestamp:   now,
+						ID:          i,
+						Name:        fmt.Sprintf("GPU %d", i),
+						TempC:       tempC,
+						GpuUtilPct:  gpuUtil,
+						MemUtilPct:  memUtilPct,
+						MemUsedMB:   memUsedMB,
+						MemTotalMB:  memTotalMB,
+						FanSpeedPct: fanSpeedPct,
+						PowerDrawW:  powerDrawW,
+					})
+				}
+
+				if len(stats) > 0 {
+					select {
+					case ch <- stats:
+					default:
+					}
+				}
+			}
+		}
+	}()
+
+	return ch, nil
+}

internal/perf/d3dkmt_windows_test.go

@@ -0,0 +1,98 @@
+//go:build windows
+
+package perf
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestD3dkmtNodeUtil_FullLoad(t *testing.T) {
+	prev := nodeRunningTimes{Global: 1000, System: 10000}
+	cur := nodeRunningTimes{Global: 5000, System: 14000}
+	got := d3dkmtNodeUtil(prev, cur, 100000)
+	assert.Equal(t, 100.0, got)
+}
+
+func TestD3dkmtNodeUtil_PartialUtil(t *testing.T) {
+	prev := nodeRunningTimes{Global: 1000, System: 10000}
+	cur := nodeRunningTimes{Global: 3000, System: 14000}
+	got := d3dkmtNodeUtil(prev, cur, 100000)
+	assert.Equal(t, 50.0, got)
+}
+
+func TestD3dkmtNodeUtil_Identical(t *testing.T) {
+	prev := nodeRunningTimes{Global: 10000, System: 10000}
+	cur := nodeRunningTimes{Global: 20000, System: 20000}
+	got := d3dkmtNodeUtil(prev, cur, 100000)
+	assert.Equal(t, 100.0, got)
+}
+
+func TestD3dkmtNodeUtil_CounterWrap(t *testing.T) {
+	prev := nodeRunningTimes{Global: 9000, System: 10000}
+	cur := nodeRunningTimes{Global: 1000, System: 10000}
+	got := d3dkmtNodeUtil(prev, cur, 100000)
+	assert.Equal(t, -1.0, got)
+}
+
+func TestD3dkmtNodeUtil_SystemWrap(t *testing.T) {
+	prev := nodeRunningTimes{Global: 1000, System: 9000}
+	cur := nodeRunningTimes{Global: 5000, System: 1000}
+	got := d3dkmtNodeUtil(prev, cur, 100000)
+	assert.Equal(t, -1.0, got)
+}
+
+func TestD3dkmtNodeUtil_ZeroDelta(t *testing.T) {
+	prev := nodeRunningTimes{Global: 1000, System: 10000}
+	cur := nodeRunningTimes{Global: 1000, System: 10000}
+	got := d3dkmtNodeUtil(prev, cur, 100000)
+	assert.Equal(t, 0.0, got)
+}
+
+func TestD3dkmtNodeUtil_ElapsedFallback(t *testing.T) {
+	prev := nodeRunningTimes{Global: 1000, System: 10000}
+	cur := nodeRunningTimes{Global: 6000, System: 10000}
+	got := d3dkmtNodeUtil(prev, cur, 50000)
+	assert.InDelta(t, 10.0, got, 0.01)
+}
+
+func TestD3dkmtFanPct_Normal(t *testing.T) {
+	assert.Equal(t, 50.0, d3dkmtFanPct(1500, 3000))
+}
+
+func TestD3dkmtFanPct_MaxFan(t *testing.T) {
+	assert.Equal(t, 100.0, d3dkmtFanPct(3000, 3000))
+}
+
+func TestD3dkmtFanPct_OverMaxClamped(t *testing.T) {
+	assert.Equal(t, 100.0, d3dkmtFanPct(4000, 3000))
+}
+
+func TestD3dkmtFanPct_ZeroMaxFan(t *testing.T) {
+	assert.Equal(t, 0.0, d3dkmtFanPct(1500, 0))
+}
+
+func TestD3dkmtFanPct_ZeroFanRPM(t *testing.T) {
+	assert.Equal(t, 0.0, d3dkmtFanPct(0, 3000))
+}
+
+func TestD3dkmtFanPct_BothZero(t *testing.T) {
+	assert.Equal(t, 0.0, d3dkmtFanPct(0, 0))
+}
+
+func TestD3dkmtPowerW(t *testing.T) {
+	assert.Equal(t, 250.0, d3dkmtPowerW(2500))
+}
+
+func TestD3dkmtPowerW_Zero(t *testing.T) {
+	assert.Equal(t, 0.0, d3dkmtPowerW(0))
+}
+
+func TestD3dkmtTempC(t *testing.T) {
+	assert.Equal(t, 65, d3dkmtTempC(650))
+}
+
+func TestD3dkmtTempC_Zero(t *testing.T) {
+	assert.Equal(t, 0, d3dkmtTempC(0))
+}

internal/perf/gpu_parse.go

@@ -1,6 +1,10 @@
 package perf
 
 import (
+	"encoding/json"
+	"fmt"
+	"math"
+	"regexp"
 	"strconv"
 	"strings"
 	"time"
@@ -43,3 +47,168 @@ func ParseNvidiaSmiLine(line string) *GpuStat {
 		PowerDrawW:  powerDraw,
 	}
 }
+
+// mactopOutput maps the subset of mactop's headless JSON output that is
+// relevant to GpuStat. Note that mactop's memory object is whole-system memory,
+// not GPU-attributed; the darwin monitor overlays ioreg's GPU-attributed
+// unified memory (see overlayIoregMem) so both backends report consistent
+// memory figures.
+type mactopOutput struct {
+	SocMetrics struct {
+		GPUPower float64 `json:"gpu_power"`
+		GPUFreq  int     `json:"gpu_freq_mhz"`
+		GPUTemp  float64 `json:"gpu_temp"`
+	} `json:"soc_metrics"`
+	Memory struct {
+		Total uint64 `json:"total"`
+		Used  uint64 `json:"used"`
+	} `json:"memory"`
+	GPUUsage   float64 `json:"gpu_usage"`
+	SystemInfo struct {
+		Name         string `json:"name"`
+		GPUCoreCount int    `json:"gpu_core_count"`
+	} `json:"system_info"`
+	Fans []struct {
+		RPM    int `json:"rpm"`
+		MinRPM int `json:"min_rpm"`
+		MaxRPM int `json:"max_rpm"`
+	} `json:"fans"`
+	Temperatures []struct {
+		Group string  `json:"group"`
+		Avg   float64 `json:"avg_celsius"`
+	} `json:"temperatures"`
+}
+
+// ioreg output uses ` = ` (with spaces) for top-level device properties and
+// `=` (no spaces) for values inside nested dictionaries such as
+// PerformanceStatistics.
+var (
+	reIoregModel     = regexp.MustCompile(`"model"\s*=\s*"([^"]+)"`)
+	reIoregCoreCount = regexp.MustCompile(`"gpu-core-count"\s*=\s*(\d+)`)
+	reIoregUtil      = regexp.MustCompile(`"Device Utilization %"=(\d+)`)
+	reIoregMemUsed   = regexp.MustCompile(`"In use system memory"=(\d+)`)
+)
+
+// ParseIoregOutput parses `ioreg -r -c IOGPU -d 1 -f` output into a GpuStat for
+// the Apple Silicon integrated GPU. This is a fallback for when mactop is not
+// installed: utilization and used memory are available, but power, temperature,
+// and fan speed are not exposed by ioreg. memTotalMB is the unified memory size
+// supplied by the caller, since Apple Silicon shares memory between CPU and GPU.
+// Returns nil if no GPU device is found in the output.
+func ParseIoregOutput(out []byte, memTotalMB int) *GpuStat {
+	utilMatch := reIoregUtil.FindSubmatch(out)
+	memMatch := reIoregMemUsed.FindSubmatch(out)
+	if utilMatch == nil && memMatch == nil {
+		return nil
+	}
+
+	var gpuUtil float64
+	if utilMatch != nil {
+		gpuUtil, _ = strconv.ParseFloat(string(utilMatch[1]), 64)
+	}
+
+	const toMB = 1024 * 1024
+	var memUsedMB int
+	if memMatch != nil {
+		memUsedBytes, _ := strconv.ParseInt(string(memMatch[1]), 10, 64)
+		memUsedMB = int(memUsedBytes / toMB)
+	}
+
+	var memUtil float64
+	if memTotalMB > 0 {
+		memUtil = float64(memUsedMB) / float64(memTotalMB) * 100
+	}
+
+	name := "Apple GPU"
+	if m := reIoregModel.FindSubmatch(out); m != nil {
+		name = string(m[1])
+	}
+	if m := reIoregCoreCount.FindSubmatch(out); m != nil {
+		if cores, err := strconv.Atoi(string(m[1])); err == nil && cores > 0 {
+			name = fmt.Sprintf("%s (%d-core GPU)", name, cores)
+		}
+	}
+
+	return &GpuStat{
+		Timestamp:  time.Now(),
+		ID:         0,
+		Name:       name,
+		GpuUtilPct: gpuUtil,
+		MemUtilPct: memUtil,
+		MemUsedMB:  memUsedMB,
+		MemTotalMB: memTotalMB,
+	}
+}
+
+// ParseMactopLine parses a single line of mactop headless JSON output into a
+// GpuStat for the Apple Silicon integrated GPU. Returns nil if the line cannot
+// be parsed.
+func ParseMactopLine(line string) *GpuStat {
+	line = strings.TrimSpace(line)
+	if line == "" {
+		return nil
+	}
+
+	var out mactopOutput
+	if err := json.Unmarshal([]byte(line), &out); err != nil {
+		return nil
+	}
+
+	const toMB = 1024 * 1024
+	memUsedMB := int(out.Memory.Used / toMB)
+	memTotalMB := int(out.Memory.Total / toMB)
+
+	var memUtil float64
+	if memTotalMB > 0 {
+		memUtil = float64(memUsedMB) / float64(memTotalMB) * 100
+	}
+
+	name := out.SystemInfo.Name
+	if name == "" {
+		name = "Apple GPU"
+	}
+	if out.SystemInfo.GPUCoreCount > 0 {
+		name = fmt.Sprintf("%s (%d-core GPU)", name, out.SystemInfo.GPUCoreCount)
+	}
+
+	// Unified memory has no dedicated VRAM sensor; use the memory temperature
+	// group when mactop exposes it.
+	var vramTempC int
+	for _, t := range out.Temperatures {
+		if strings.EqualFold(t.Group, "Memory") {
+			vramTempC = int(math.Round(t.Avg))
+			break
+		}
+	}
+
+	// Average fan load across all fans as a percentage of their RPM range.
+	var fanSpeed float64
+	var fanCount int
+	for _, f := range out.Fans {
+		if f.MaxRPM > f.MinRPM {
+			pct := float64(f.RPM-f.MinRPM) / float64(f.MaxRPM-f.MinRPM) * 100
+			if pct < 0 {
+				pct = 0
+			}
+			fanSpeed += pct
+			fanCount++
+		}
+	}
+	if fanCount > 0 {
+		fanSpeed /= float64(fanCount)
+	}
+
+	return &GpuStat{
+		Timestamp:   time.Now(),
+		ID:          0,
+		Name:        name,
+		TempC:       int(math.Round(out.SocMetrics.GPUTemp)),
+		VramTempC:   vramTempC,
+		GpuUtilPct:  out.GPUUsage,
+		MemUtilPct:  memUtil,
+		MemUsedMB:   memUsedMB,
+		MemTotalMB:  memTotalMB,
+		FanSpeedPct: fanSpeed,
+		PowerDrawW:  out.SocMetrics.GPUPower,
+	}
+}

internal/perf/monitor.go

@@ -12,7 +12,7 @@ import (
 )
 
 var (
-	ErrNotImplemented = errors.New("Not Implemented")
+	ErrNotImplemented = errors.New("not implemented")
 	ErrNoGpuTool      = errors.New("no GPU monitoring tool available")
 )
 

internal/perf/monitor_darwin.go

@@ -1,7 +1,11 @@
 package perf
 
 import (
+	"bufio"
 	"context"
+	"fmt"
+	"os/exec"
+	"strings"
 	"time"
 
 	"github.com/mostlygeek/llama-swap/internal/logmon"
@@ -11,7 +15,156 @@ import (
 )
 
 func getGpuStats(ctx context.Context, every time.Duration, logger *logmon.Monitor) (chan []GpuStat, error) {
-	return nil, ErrNotImplemented
+	if ch, err := tryMactop(ctx, every, logger); err == nil {
+		logger.Info("using mactop for GPU monitoring")
+		return ch, nil
+	} else {
+		logger.Debugf("mactop: %s", err.Error())
+	}
+
+	if ch, err := tryIoreg(ctx, every, logger); err == nil {
+		logger.Info("using ioreg for GPU monitoring")
+		return ch, nil
+	} else {
+		logger.Debugf("ioreg: %s", err.Error())
+	}
+
+	return nil, ErrNoGpuTool
+}
+
+// tryIoreg polls `ioreg -r -c IOGPU -d 1 -f` for Apple Silicon GPU stats. It is
+// a fallback for when mactop is not installed. ioreg exposes GPU utilization and
+// used memory but not power, temperature, or fan speed.
+func tryIoreg(ctx context.Context, every time.Duration, logger *logmon.Monitor) (chan []GpuStat, error) {
+	if _, err := exec.LookPath("ioreg"); err != nil {
+		return nil, ErrNoGpuTool
+	}
+
+	// Verify ioreg actually reports a GPU device before committing to it, so we
+	// can fall through to ErrNoGpuTool otherwise.
+	if stat := sampleIoreg(ctx); stat == nil {
+		return nil, fmt.Errorf("ioreg reported no GPU device")
+	}
+
+	if every < time.Second {
+		every = time.Second
+	}
+
+	ch := make(chan []GpuStat, 1)
+
+	go func() {
+		defer close(ch)
+		ticker := time.NewTicker(every)
+		defer ticker.Stop()
+
+		for {
+			select {
+			case <-ctx.Done():
+				return
+			case <-ticker.C:
+				stat := sampleIoreg(ctx)
+				if stat == nil {
+					continue
+				}
+				select {
+				case ch <- []GpuStat{*stat}:
+				default:
+				}
+			}
+		}
+	}()
+
+	return ch, nil
+}
+
+// sampleIoreg runs ioreg once and parses a single GpuStat, or returns nil.
+func sampleIoreg(ctx context.Context) *GpuStat {
+	out, err := exec.CommandContext(ctx, "ioreg", "-r", "-c", "IOGPU", "-d", "1", "-f").Output()
+	if err != nil {
+		return nil
+	}
+
+	var memTotalMB int
+	if vmStat, err := mem.VirtualMemory(); err == nil {
+		memTotalMB = int(vmStat.Total / (1024 * 1024))
+	}
+
+	return ParseIoregOutput(out, memTotalMB)
+}
+
+// overlayIoregMem replaces a GpuStat's memory fields with the GPU-attributed
+// unified memory reported by ioreg. mactop only exposes whole-system memory, so
+// without this the mactop and ioreg backends would report different memory
+// semantics. It is a no-op when ioreg is unavailable or reports no GPU memory,
+// leaving the mactop-supplied values in place.
+func overlayIoregMem(ctx context.Context, stat *GpuStat) {
+	ioStat := sampleIoreg(ctx)
+	if ioStat == nil {
+		return
+	}
+	stat.MemUsedMB = ioStat.MemUsedMB
+	stat.MemTotalMB = ioStat.MemTotalMB
+	stat.MemUtilPct = ioStat.MemUtilPct
+}
+
+// tryMactop streams Apple Silicon GPU stats from mactop's headless mode.
+// See https://github.com/metaspartan/mactop. mactop emits one JSON object per
+// sample to stdout, which we parse into GpuStat.
+func tryMactop(ctx context.Context, every time.Duration, logger *logmon.Monitor) (chan []GpuStat, error) {
+	if _, err := exec.LookPath("mactop"); err != nil {
+		return nil, ErrNoGpuTool
+	}
+
+	// mactop samples power over the interval, so give it at least a second.
+	intervalMs := int(every.Milliseconds())
+	if intervalMs < 1000 {
+		intervalMs = 1000
+	}
+
+	cmd := exec.CommandContext(ctx, "mactop",
+		"--headless",
+		"--format", "json",
+		"--interval", fmt.Sprintf("%d", intervalMs),
+	)
+
+	stdout, err := cmd.StdoutPipe()
+	if err != nil {
+		return nil, fmt.Errorf("mactop stdout pipe failed: %w", err)
+	}
+
+	if err := cmd.Start(); err != nil {
+		return nil, fmt.Errorf("mactop start failed: %w", err)
+	}
+
+	ch := make(chan []GpuStat, 1)
+
+	go func() {
+		defer close(ch)
+
+		scanner := bufio.NewScanner(stdout)
+		// mactop's JSON objects can be large; allow generous line lengths.
+		scanner.Buffer(make([]byte, 0, 64*1024), 1024*1024)
+		for scanner.Scan() {
+			line := strings.TrimSpace(scanner.Text())
+			if line == "" {
+				continue
+			}
+
+			stat := ParseMactopLine(line)
+			if stat != nil {
+				// mactop only reports whole-system memory; overlay ioreg's
+				// GPU-attributed unified memory so both backends are consistent.
+				overlayIoregMem(ctx, stat)
+				select {
+				case ch <- []GpuStat{*stat}:
+				default:
+				}
+			}
+		}
+		cmd.Wait()
+	}()
+
+	return ch, nil
 }
 
 func readSysStats() (SysStat, error) {

internal/perf/monitor_test.go

@@ -264,3 +264,50 @@ func TestParseNvidiaSmiLine_ZeroMemoryTotal(t *testing.T) {
 	require.NotNil(t, stat)
 	assert.Equal(t, 0.0, stat.MemUtilPct)
 }
+
+const ioregSample = `+-o AGXAcceleratorG13X  <class AGXAcceleratorG13X, id 0x1000009a1, registered, matched, active, busy 0 (39191 ms), retain 108>
+    {
+      "model" = "Apple M1 Pro"
+      "gpu-core-count" = 16
+      "PerformanceStatistics" = {"In use system memory (driver)"=0,"Alloc system memory"=14511046656,"Tiler Utilization %"=34,"recoveryCount"=0,"Renderer Utilization %"=34,"Device Utilization %"=34,"In use system memory"=7688503296}
+      "IOClass" = "AGXAcceleratorG13X"
+    }`
+
+func TestParseIoregOutput_ValidOutput(t *testing.T) {
+	const memTotalMB = 32768
+
+	stat := ParseIoregOutput([]byte(ioregSample), memTotalMB)
+	require.NotNil(t, stat)
+
+	assert.Equal(t, 0, stat.ID)
+	assert.Equal(t, "Apple M1 Pro (16-core GPU)", stat.Name)
+	assert.Equal(t, 34.0, stat.GpuUtilPct)
+	assert.Equal(t, 7688503296/(1024*1024), stat.MemUsedMB)
+	assert.Equal(t, memTotalMB, stat.MemTotalMB)
+	assert.InDelta(t, float64(stat.MemUsedMB)/memTotalMB*100, stat.MemUtilPct, 0.01)
+	// Not exposed by ioreg.
+	assert.Equal(t, 0, stat.TempC)
+	assert.Equal(t, 0.0, stat.PowerDrawW)
+	assert.Equal(t, 0.0, stat.FanSpeedPct)
+}
+
+func TestParseIoregOutput_NoGpuDevice(t *testing.T) {
+	stat := ParseIoregOutput([]byte("no gpu here"), 32768)
+	assert.Nil(t, stat)
+}
+
+func TestParseIoregOutput_ZeroMemTotal(t *testing.T) {
+	stat := ParseIoregOutput([]byte(ioregSample), 0)
+	require.NotNil(t, stat)
+	assert.Equal(t, 0.0, stat.MemUtilPct)
+}
+
+func TestParseIoregOutput_MissingModel(t *testing.T) {
+	const out = `"Device Utilization %"=50,"In use system memory"=1048576`
+
+	stat := ParseIoregOutput([]byte(out), 1024)
+	require.NotNil(t, stat)
+	assert.Equal(t, "Apple GPU", stat.Name)
+	assert.Equal(t, 50.0, stat.GpuUtilPct)
+	assert.Equal(t, 1, stat.MemUsedMB)
+}

internal/perf/monitor_windows.go

@@ -22,6 +22,13 @@ func getGpuStats(ctx context.Context, every time.Duration, logger *logmon.Monito
 		logger.Debugf("nvidia-smi: %s", err.Error())
 	}
 
+	if ch, err := tryD3DKMT(ctx, every, logger); err == nil {
+		logger.Info("using D3DKMT for GPU monitoring")
+		return ch, nil
+	} else {
+		logger.Debugf("D3DKMT: %s", err.Error())
+	}
+
 	return nil, ErrNoGpuTool
 }
 

internal/perf/pdh_windows.go

@@ -0,0 +1,159 @@
+//go:build windows
+
+package perf
+
+import (
+	"fmt"
+	"strconv"
+	"strings"
+	"unsafe"
+
+	"golang.org/x/sys/windows"
+)
+
+var (
+	pdhDLL                          = windows.NewLazySystemDLL("pdh.dll")
+	procPdhOpenQuery                = pdhDLL.NewProc("PdhOpenQueryW")
+	procPdhAddEnglishCounter        = pdhDLL.NewProc("PdhAddEnglishCounterW")
+	procPdhCollectQueryData         = pdhDLL.NewProc("PdhCollectQueryData")
+	procPdhGetFormattedCounterArray = pdhDLL.NewProc("PdhGetFormattedCounterArrayW")
+	procPdhCloseQuery               = pdhDLL.NewProc("PdhCloseQuery")
+)
+
+const (
+	pdhFmtDouble = 0x00000200
+	pdhMoreData  = 0x800007D2
+	pdhNoData    = 0x800007D5
+)
+
+type pdhCounterValue struct {
+	CStatus uint32
+	DblVal  float64
+}
+
+type pdhCounterValueItem struct {
+	SzName   *uint16
+	FmtValue pdhCounterValue
+}
+
+func init() {
+	var item pdhCounterValueItem
+	if unsafe.Sizeof(item) != 24 {
+		panic(fmt.Sprintf("pdhCounterValueItem size %d != expected 24 on x64", unsafe.Sizeof(item)))
+	}
+}
+
+type pdhGpuUtil struct {
+	query   uintptr
+	counter uintptr
+}
+
+// initPdhGpuUtil creates a PDH query for the GPU Engine utilization counter.
+// Returns nil with an error if PDH or the counter is unavailable.
+func initPdhGpuUtil() (*pdhGpuUtil, error) {
+	var query uintptr
+	if ret, _, _ := procPdhOpenQuery.Call(0, 0, uintptr(unsafe.Pointer(&query))); ret != 0 {
+		return nil, fmt.Errorf("PdhOpenQuery: 0x%x", ret)
+	}
+
+	path, _ := windows.UTF16PtrFromString(`\GPU Engine(*)\Utilization Percentage`)
+	var counter uintptr
+	if ret, _, _ := procPdhAddEnglishCounter.Call(
+		query, uintptr(unsafe.Pointer(path)), 0, uintptr(unsafe.Pointer(&counter)),
+	); ret != 0 {
+		procPdhCloseQuery.Call(query)
+		return nil, fmt.Errorf("PdhAddEnglishCounter(GPU Engine): 0x%x", ret)
+	}
+
+	procPdhCollectQueryData.Call(query)
+
+	return &pdhGpuUtil{query: query, counter: counter}, nil
+}
+
+// close releases the PDH query handle.
+func (p *pdhGpuUtil) close() {
+	if p.query != 0 {
+		procPdhCloseQuery.Call(p.query)
+		p.query = 0
+	}
+}
+
+// collect reads the PDH counter and returns a map of adapter LUID to
+// aggregated GPU utilization percentage, summed across all engine instances
+// per adapter and clamped to 100%.
+func (p *pdhGpuUtil) collect() map[LUID]float64 {
+	ret, _, _ := procPdhCollectQueryData.Call(p.query)
+	if ret != 0 && ret != pdhNoData {
+		return nil
+	}
+
+	var bufSize uint32
+	var itemCount uint32
+	ret, _, _ = procPdhGetFormattedCounterArray.Call(
+		p.counter, pdhFmtDouble,
+		uintptr(unsafe.Pointer(&bufSize)),
+		uintptr(unsafe.Pointer(&itemCount)),
+		0,
+	)
+	if ret != pdhMoreData || itemCount == 0 {
+		return nil
+	}
+
+	buf := make([]byte, bufSize)
+	ret, _, _ = procPdhGetFormattedCounterArray.Call(
+		p.counter, pdhFmtDouble,
+		uintptr(unsafe.Pointer(&bufSize)),
+		uintptr(unsafe.Pointer(&itemCount)),
+		uintptr(unsafe.Pointer(&buf[0])),
+	)
+	if ret != 0 {
+		return nil
+	}
+
+	itemSize := uint32(unsafe.Sizeof(pdhCounterValueItem{}))
+	result := make(map[LUID]float64)
+
+	for i := uint32(0); i < itemCount; i++ {
+		item := (*pdhCounterValueItem)(unsafe.Pointer(&buf[i*itemSize]))
+		if item.FmtValue.CStatus != 0 {
+			continue
+		}
+		luid, ok := parsePdhLuid(windows.UTF16PtrToString(item.SzName))
+		if !ok {
+			continue
+		}
+		result[luid] += item.FmtValue.DblVal
+	}
+
+	for luid := range result {
+		if result[luid] > 100.0 {
+			result[luid] = 100.0
+		}
+	}
+
+	return result
+}
+
+// parsePdhLuid extracts the adapter LUID (high and low parts) from a PDH
+// GPU Engine instance name (e.g. "pid_1234_luid_0x00000000_0x000148BF_phys_0_eng_2_engtype_Compute").
+func parsePdhLuid(name string) (LUID, bool) {
+	idx := strings.Index(name, "luid_0x")
+	if idx < 0 {
+		return LUID{}, false
+	}
+	rest := name[idx+7:]
+	parts := strings.SplitN(rest, "_", 4)
+	if len(parts) < 3 {
+		return LUID{}, false
+	}
+	hp, err := strconv.ParseUint(parts[0], 16, 32)
+	if err != nil {
+		return LUID{}, false
+	}
+	lpStr := strings.TrimPrefix(parts[1], "0x")
+	lp, err := strconv.ParseUint(lpStr, 16, 32)
+	if err != nil {
+		return LUID{}, false
+	}
+	return LUID{LowPart: uint32(lp), HighPart: int32(hp)}, true
+}

internal/perf/pdh_windows_test.go

@@ -0,0 +1,53 @@
+//go:build windows
+
+package perf
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestParsePdhLuid_Valid(t *testing.T) {
+	name := `pid_25312_luid_0x00000000_0x000148BF_phys_0_eng_2_engtype_Compute`
+	got, ok := parsePdhLuid(name)
+	assert.True(t, ok)
+	assert.Equal(t, uint32(0x000148BF), got.LowPart)
+	assert.Equal(t, int32(0x00000000), got.HighPart)
+}
+
+func TestParsePdhLuid_ValidNvidia(t *testing.T) {
+	name := `pid_1388_luid_0x00000000_0x00011372_phys_0_eng_8_engtype_Compute_1`
+	got, ok := parsePdhLuid(name)
+	assert.True(t, ok)
+	assert.Equal(t, uint32(0x00011372), got.LowPart)
+	assert.Equal(t, int32(0x00000000), got.HighPart)
+}
+
+func TestParsePdhLuid_NonZeroHighPart(t *testing.T) {
+	name := `pid_1234_luid_0x00000001_0x0000C85A_phys_0_eng_5_engtype_Copy`
+	got, ok := parsePdhLuid(name)
+	assert.True(t, ok)
+	assert.Equal(t, uint32(0x0000C85A), got.LowPart)
+	assert.Equal(t, int32(0x00000001), got.HighPart)
+}
+
+func TestParsePdhLuid_InvalidNoLuid(t *testing.T) {
+	_, ok := parsePdhLuid("invalid_string_without_luid")
+	assert.False(t, ok)
+}
+
+func TestParsePdhLuid_InvalidEmpty(t *testing.T) {
+	_, ok := parsePdhLuid("")
+	assert.False(t, ok)
+}
+
+func TestParsePdhLuid_InvalidHex(t *testing.T) {
+	_, ok := parsePdhLuid("pid_1234_luid_0xZZZZ_0xGGGG_phys_0")
+	assert.False(t, ok)
+}
+
+func TestParsePdhLuid_ShortAfterLuid(t *testing.T) {
+	_, ok := parsePdhLuid("pid_1234_luid_0x00000000")
+	assert.False(t, ok)
+}

internal/process/process_command.go

@@ -11,7 +11,6 @@ import (
 	"os/exec"
 	"strings"
 	"sync/atomic"
-	"syscall"
 	"time"
 
 	"github.com/mostlygeek/llama-swap/internal/config"
@@ -22,6 +21,22 @@ import (
 
 var ErrStartAborted = fmt.Errorf("aborted")
 
+// cmdWaitDelay is the upper bound the runtime will wait for child I/O to
+// drain after the process exits before force-closing the stdout/stderr
+// pipes. Required so that cmd.Wait() returns even when a forked grandchild
+// inherits and holds the pipes open (e.g. a shell wrapper that backgrounds
+// the real binary). killProcess sends the stop signal directly (not via the
+// cmd context), so this delay is measured from process exit rather than from
+// the stop request, and stays independent of the caller's graceful timeout.
+const cmdWaitDelay = 10 * time.Second
+
+// parentCancelGraceTimeout is the graceful timeout used when the process is
+// torn down because parentCtx was cancelled (final router teardown or app
+// shutdown). In the normal flow the process has already been stopped via
+// Stop() by this point, so killProcess is a no-op kill; the short grace just
+// bounds the rare case where a process is still alive when its context is cut.
+const parentCancelGraceTimeout = time.Second
+
 type runReq struct {
 	timeout time.Duration
 	respond chan error
@@ -39,6 +54,7 @@ type waitReadyReq struct {
 type startResult struct {
 	cmd       *exec.Cmd
 	cmdDone   chan struct{}
+	cancel    context.CancelFunc
 	handlerFn http.HandlerFunc
 	err       error
 }
@@ -51,6 +67,11 @@ type ProcessCommand struct {
 	processLogger *logmon.Monitor
 	proxyLogger   *logmon.Monitor
 
+	// waitDelay is assigned to cmd.WaitDelay when starting the upstream
+	// process. Defaults to cmdWaitDelay; tests override it to keep the
+	// pipe-close backstop from dominating their runtime.
+	waitDelay time.Duration
+
 	runCh       chan runReq
 	stopCh      chan stopReq
 	waitReadyCh chan waitReadyReq
@@ -85,6 +106,7 @@ func New(
 		runCh:       make(chan runReq),
 		stopCh:      make(chan stopReq),
 		waitReadyCh: make(chan waitReadyReq),
+		waitDelay:   cmdWaitDelay,
 	}
 	p.state.Store(StateStopped)
 
@@ -122,6 +144,7 @@ func (p *ProcessCommand) run() {
 	var (
 		cmd          *exec.Cmd
 		cmdDone      <-chan struct{}
+		cmdCancel    context.CancelFunc
 		readyWaiters []waitReadyReq
 		// runResp parks the in-flight Run caller's response channel. The
 		// interface contract is that Run blocks until the process is
@@ -164,9 +187,10 @@ func (p *ProcessCommand) run() {
 			setState(StateShutdown)
 			if cmd != nil {
 				p.handler.Store(nil)
-				p.killProcess(cmd, cmdDone, 100*time.Millisecond)
+				p.killProcess(cmd, cmdCancel, cmdDone, parentCancelGraceTimeout)
 				cmd = nil
 				cmdDone = nil
+				cmdCancel = nil
 			}
 			notifyWaiters(fmt.Errorf("[%s] shutdown", p.id))
 			respondRun(fmt.Errorf("[%s] shutdown", p.id))
@@ -177,8 +201,12 @@ func (p *ProcessCommand) run() {
 		// cmdDone is nil while no process is running, so this case is
 		// dormant outside of StateReady.
 		case <-cmdDone:
+			if cmdCancel != nil {
+				cmdCancel()
+			}
 			cmd = nil
 			cmdDone = nil
+			cmdCancel = nil
 			p.handler.Store(nil)
 			setState(StateStopped)
 			respondRun(fmt.Errorf("[%s] upstream exited unexpectedly", p.id))
@@ -226,6 +254,7 @@ func (p *ProcessCommand) run() {
 				if res.err == nil {
 					cmd = res.cmd
 					cmdDone = res.cmdDone
+					cmdCancel = res.cancel
 					fn := res.handlerFn
 					p.handler.Store(&fn)
 					setState(StateReady)
@@ -273,7 +302,7 @@ func (p *ProcessCommand) run() {
 				cancelStart()
 				res := <-resultCh
 				if res.cmd != nil {
-					p.killProcess(res.cmd, res.cmdDone, stop.timeout)
+					p.killProcess(res.cmd, res.cancel, res.cmdDone, stop.timeout)
 				}
 				setState(StateStopped)
 				notifyWaiters(ErrStartAborted)
@@ -293,7 +322,7 @@ func (p *ProcessCommand) run() {
 				setState(StateShutdown)
 				res := <-resultCh
 				if res.cmd != nil {
-					p.killProcess(res.cmd, res.cmdDone, 100*time.Millisecond)
+					p.killProcess(res.cmd, res.cancel, res.cmdDone, parentCancelGraceTimeout)
 				}
 				notifyWaiters(fmt.Errorf("[%s] shutdown", p.id))
 				respondRun(fmt.Errorf("[%s] shutdown", p.id))
@@ -310,9 +339,10 @@ func (p *ProcessCommand) run() {
 		case stop := <-p.stopCh:
 			if cmd != nil {
 				setState(StateStopping)
-				p.killProcess(cmd, cmdDone, stop.timeout)
+				p.killProcess(cmd, cmdCancel, cmdDone, stop.timeout)
 				cmd = nil
 				cmdDone = nil
+				cmdCancel = nil
 				p.handler.Store(nil)
 			}
 			// Stop is a no-op (and not an error) when already Stopped — this
@@ -377,46 +407,71 @@ func (p *ProcessCommand) doStart(startCtx context.Context, healthCheckTimeout ti
 		reverseProxy.ServeHTTP(w, r)
 	})
 
-	cmd := exec.Command(args[0], args[1:]...)
+	// cmdCtx + cmd.Cancel are wired as a safety net: if the context is ever
+	// cancelled while the process is alive, cmd.Cancel sends SIGTERM / CmdStop
+	// and the runtime escalates to SIGKILL after cmd.WaitDelay. In the normal
+	// teardown path killProcess sends the stop signal directly instead, so
+	// cmd.WaitDelay only acts as the inherited-pipe backstop measured from
+	// process exit (see killProcess).
+	cmdCtx, cmdCancel := context.WithCancel(context.Background())
+	cmd := exec.CommandContext(cmdCtx, args[0], args[1:]...)
 	cmd.Stderr = p.processLogger
 	cmd.Stdout = p.processLogger
 	cmd.Env = append(cmd.Environ(), p.config.Env...)
+	cmd.Cancel = func() error { return p.sendStopSignal(cmd) }
+	cmd.WaitDelay = p.waitDelay
 	setProcAttributes(cmd)
 
 	p.proxyLogger.Debugf("<%s> Executing start command: %s, env: %s", p.id, strings.Join(args, " "), strings.Join(p.config.Env, ", "))
 
 	cmdDone := make(chan struct{})
 	if err := cmd.Start(); err != nil {
+		cmdCancel()
 		return startResult{err: fmt.Errorf("failed to start command '%s': %w", strings.Join(args, " "), err)}
 	}
 
 	go func() {
 		waitErr := cmd.Wait()
-		if exitErr, ok := waitErr.(*exec.ExitError); ok {
-			p.proxyLogger.Debugf("<%s> process exited: code=%d, err=%v", p.id, exitErr.ExitCode(), waitErr)
-		} else if waitErr != nil {
-			p.proxyLogger.Debugf("<%s> process exited with error: %v", p.id, waitErr)
-		} else {
+		switch st := p.State(); {
+		case waitErr == nil:
 			p.proxyLogger.Debugf("<%s> process exited cleanly", p.id)
+		case st == StateStopping || st == StateShutdown:
+			// Expected: we force-terminated the process. A forced kill exits
+			// the child with a non-zero code (e.g. taskkill /f on Windows
+			// yields exit status 1), so this is not an error.
+			p.proxyLogger.Debugf("<%s> process stopped by llama-swap: %v", p.id, waitErr)
+		default:
+			if exitErr, ok := waitErr.(*exec.ExitError); ok {
+				p.proxyLogger.Debugf("<%s> process exited: code=%d, err=%v", p.id, exitErr.ExitCode(), waitErr)
+			} else {
+				p.proxyLogger.Debugf("<%s> process exited with error: %v", p.id, waitErr)
+			}
 		}
 		close(cmdDone)
 	}()
 
+	abort := func(err error) startResult {
+		p.killProcess(cmd, cmdCancel, cmdDone, 5*time.Second)
+		return startResult{err: err}
+	}
+	prematureExit := func() startResult {
+		cmdCancel()
+		return startResult{err: fmt.Errorf("upstream command exited prematurely")}
+	}
+
 	if startCtx.Err() != nil {
-		p.killProcess(cmd, cmdDone, 5*time.Second)
-		return startResult{err: ErrStartAborted}
+		return abort(ErrStartAborted)
 	}
 
 	checkEndpoint := strings.TrimSpace(p.config.CheckEndpoint)
 	if checkEndpoint == "none" {
-		return startResult{cmd: cmd, cmdDone: cmdDone, handlerFn: handlerFn}
+		return startResult{cmd: cmd, cmdDone: cmdDone, cancel: cmdCancel, handlerFn: handlerFn}
 	}
 
 	// Wait 250ms for the command to start up before health checking
 	select {
 	case <-startCtx.Done():
-		p.killProcess(cmd, cmdDone, 5*time.Second)
-		return startResult{err: ErrStartAborted}
+		return abort(ErrStartAborted)
 	case <-time.After(250 * time.Millisecond):
 	}
 
@@ -424,16 +479,14 @@ func (p *ProcessCommand) doStart(startCtx context.Context, healthCheckTimeout ti
 	for {
 		select {
 		case <-startCtx.Done():
-			p.killProcess(cmd, cmdDone, 5*time.Second)
-			return startResult{err: ErrStartAborted}
+			return abort(ErrStartAborted)
 		case <-cmdDone:
-			return startResult{err: fmt.Errorf("upstream command exited prematurely")}
+			return prematureExit()
 		default:
 		}
 
 		if time.Now().After(deadline) {
-			p.killProcess(cmd, cmdDone, 5*time.Second)
-			return startResult{err: fmt.Errorf("health check timed out after %v", healthCheckTimeout)}
+			return abort(fmt.Errorf("health check timed out after %v", healthCheckTimeout))
 		}
 
 		req, _ := http.NewRequestWithContext(startCtx, "GET", p.config.CheckEndpoint, nil)
@@ -445,42 +498,99 @@ func (p *ProcessCommand) doStart(startCtx context.Context, healthCheckTimeout ti
 			p.proxyLogger.Infof("<%s> Health check passed on %s%s", p.id, p.config.Proxy, p.config.CheckEndpoint)
 			break
 		} else if startCtx.Err() != nil {
-			p.killProcess(cmd, cmdDone, 5*time.Second)
-			return startResult{err: ErrStartAborted}
+			return abort(ErrStartAborted)
 		}
 
 		select {
 		case <-startCtx.Done():
-			p.killProcess(cmd, cmdDone, 5*time.Second)
-			return startResult{err: ErrStartAborted}
+			return abort(ErrStartAborted)
 		case <-cmdDone:
-			return startResult{err: fmt.Errorf("upstream command exited prematurely")}
+			return prematureExit()
 		case <-time.After(time.Second):
 		}
 	}
 
-	return startResult{cmd: cmd, cmdDone: cmdDone, handlerFn: handlerFn}
+	return startResult{cmd: cmd, cmdDone: cmdDone, cancel: cmdCancel, handlerFn: handlerFn}
 }
 
-func (p *ProcessCommand) killProcess(cmd *exec.Cmd, cmdDone <-chan struct{}, gracefulTimeout time.Duration) {
+// sendStopSignal runs the configured CmdStop (if any) or sends SIGTERM to
+// the upstream process. Wired up as cmd.Cancel so it fires whenever the
+// cmd's context is cancelled.
+func (p *ProcessCommand) sendStopSignal(cmd *exec.Cmd) error {
 	if cmd == nil || cmd.Process == nil {
-		return
+		p.processLogger.Debugf("<%s> sendStopSignal() called with nil cmd or process, nothing to stop", p.id)
+		return nil
 	}
-
+	pid := cmd.Process.Pid
 	if p.config.CmdStop != "" {
+		p.processLogger.Debugf("<%s> sendStopSignal() using CmdStop %q for pid %d", p.id, p.config.CmdStop, pid)
 		stopArgs, err := config.SanitizeCommand(
-			strings.ReplaceAll(p.config.CmdStop, "${PID}", fmt.Sprintf("%d", cmd.Process.Pid)),
+			strings.ReplaceAll(p.config.CmdStop, "${PID}", fmt.Sprintf("%d", pid)),
 		)
 		if err == nil {
+			p.processLogger.Debugf("<%s> sendStopSignal() running stop command: %s", p.id, strings.Join(stopArgs, " "))
 			stopCmd := exec.Command(stopArgs[0], stopArgs[1:]...)
 			stopCmd.Env = cmd.Env
 			setProcAttributes(stopCmd)
-			stopCmd.Run()
-		} else {
-			cmd.Process.Signal(syscall.SIGTERM)
+			runErr := stopCmd.Run()
+			if runErr != nil {
+				p.processLogger.Errorf("<%s> sendStopSignal() stop command failed: %v", p.id, runErr)
+			} else {
+				p.processLogger.Debugf("<%s> sendStopSignal() stop command completed for pid %d", p.id, pid)
+			}
+			return runErr
 		}
-	} else {
-		cmd.Process.Signal(syscall.SIGTERM)
+		// fall through to SIGTERM if sanitize failed
+		p.processLogger.Errorf("<%s> sendStopSignal() failed to sanitize CmdStop %q: %v, falling back to terminateProcessTree", p.id, p.config.CmdStop, err)
+	}
+	// On Unix this SIGTERMs the whole process group so a forked grandchild
+	// (e.g. a shell wrapper that backgrounds the real binary) is taken down
+	// with the parent rather than orphaned.
+	p.processLogger.Debugf("<%s> sendStopSignal() no CmdStop configured, calling terminateProcessTree for pid %d", p.id, pid)
+	termErr := terminateProcessTree(cmd)
+	if termErr != nil {
+		p.processLogger.Errorf("<%s> sendStopSignal() terminateProcessTree failed for pid %d: %v", p.id, pid, termErr)
+	}
+	return termErr
+}
+
+// killProcess terminates the upstream process. The flow:
+//
+//  1. Send the graceful stop signal (CmdStop / SIGTERM) directly — NOT by
+//     cancelling cmdCtx. Cancelling the context would start cmd.WaitDelay
+//     immediately, which force-kills the process WaitDelay after the signal
+//     and would silently cap gracefulTimeout at WaitDelay whenever
+//     gracefulTimeout is the longer of the two.
+//  2. We wait up to gracefulTimeout for the process to exit on its own.
+//  3. If still alive, we SIGKILL the process group directly (Unix) so any
+//     forked descendant is force-terminated alongside the parent.
+//  4. We wait on cmdDone. cmd.WaitDelay (set when the cmd was built) is the
+//     critical backstop here: once the process exits, if a forked grandchild
+//     inherited the stdout/stderr pipes and is still holding them, the runtime
+//     force-closes the pipes WaitDelay after the exit and cmd.Wait() unblocks.
+//     Because we never cancelled the context, that WaitDelay timer measures
+//     from process exit (see os/exec awaitGoroutines), not from this call.
+//     Without WaitDelay this select would hang forever (the v219 bug).
+//
+// cancel() is still invoked (deferred) to release the context, but only after
+// the process has exited and os/exec's ctx watcher has already torn down, so it
+// never re-fires cmd.Cancel.
+func (p *ProcessCommand) killProcess(cmd *exec.Cmd, cancel context.CancelFunc, cmdDone <-chan struct{}, gracefulTimeout time.Duration) {
+	if cancel == nil {
+		return
+	}
+	defer cancel()
+
+	// Deliver CmdStop / SIGTERM in a goroutine so a slow or hanging CmdStop
+	// cannot block the run() goroutine; the gracefulTimeout + Process.Kill
+	// path below still guarantees teardown.
+	if cmd != nil {
+		go func() {
+			p.proxyLogger.Debugf("[%s] sending stop signal with timeout %v", p.id, gracefulTimeout)
+			if err := p.sendStopSignal(cmd); err != nil {
+				p.proxyLogger.Warnf("[%s] stop signal failed: %v", p.id, err)
+			}
+		}()
 	}
 
 	timer := time.NewTimer(gracefulTimeout)
@@ -488,10 +598,16 @@ func (p *ProcessCommand) killProcess(cmd *exec.Cmd, cmdDone <-chan struct{}, gra
 
 	select {
 	case <-cmdDone:
+		return
 	case <-timer.C:
-		cmd.Process.Kill()
-		<-cmdDone
 	}
+
+	if cmd != nil {
+		// SIGKILL the whole process group on Unix so any descendant that
+		// ignored or outlived the graceful signal is force-terminated too.
+		_ = killProcessTree(cmd)
+	}
+	<-cmdDone
 }
 
 func (p *ProcessCommand) ID() string {

internal/process/process_command_forking_test.go

@@ -0,0 +1,262 @@
+//go:build !windows
+
+package process
+
+import (
+	"fmt"
+	"os"
+	"path/filepath"
+	"strconv"
+	"strings"
+	"syscall"
+	"testing"
+	"time"
+
+	"github.com/mostlygeek/llama-swap/internal/config"
+)
+
+// TestProcessCommand_StopForkingWrapper is a regression for the bug reported
+// against v219 where Stop would hang indefinitely when the upstream command
+// is a shell wrapper that forks the real binary (e.g. `#!/bin/bash` then
+// `"$@"`). After SIGTERM the wrapper dies but the grandchild inherits the
+// stdout/stderr pipes; cmd.Wait() blocks waiting for the pipe-copy goroutine
+// to drain EOF, which never happens while the grandchild holds the fds.
+//
+// The fix is cmd.WaitDelay (combined with exec.CommandContext + cmd.Cancel),
+// which causes the runtime to force-close the pipes after the delay so
+// cmd.Wait() — and therefore Stop — returns.
+func TestProcessCommand_StopForkingWrapper(t *testing.T) {
+	skipIfNoSimpleResponder(t)
+
+	port := getFreePort(t)
+	dir := t.TempDir()
+	pidFile := filepath.Join(dir, "child.pid")
+
+	// Wrapper script: backgrounds the child (which inherits stdout/stderr),
+	// records its PID for cleanup, then waits. When SIGTERM hits bash it
+	// dies without forwarding the signal; the grandchild keeps running and
+	// keeps the inherited pipe fds open. This is the scenario reported in
+	// the v219 regression.
+	wrapper := filepath.Join(dir, "wrapper.sh")
+	script := fmt.Sprintf("#!/bin/bash\n%q -port %d -silent &\necho $! > %q\nwait\n",
+		simpleResponderPath, port, pidFile)
+	if err := os.WriteFile(wrapper, []byte(script), 0o755); err != nil {
+		t.Fatalf("WriteFile: %v", err)
+	}
+	t.Cleanup(func() { killChildFromPidFile(pidFile) })
+
+	p := newProcessCommand(t, config.ModelConfig{
+		Cmd:                wrapper,
+		Proxy:              fmt.Sprintf("http://127.0.0.1:%d", port),
+		CheckEndpoint:      "/health",
+		HealthCheckTimeout: 10,
+	})
+	// Shrink the pipe-close backstop so the test doesn't sit at the
+	// production default (10s). Must be set before Run() so doStart picks
+	// it up when building the cmd.
+	const testWaitDelay = 250 * time.Millisecond
+	p.waitDelay = testWaitDelay
+
+	runErr := runAsync(t, p)
+
+	// Stop must return within a bounded time even though the grandchild
+	// is still holding the pipe open. Budget is generous on top of
+	// testWaitDelay to absorb scheduling jitter on slow CI runners; the
+	// pre-fix behaviour was an unbounded hang, so any reasonable cap
+	// distinguishes pass from fail.
+	stopReturned := make(chan error, 1)
+	stopStart := time.Now()
+	go func() { stopReturned <- p.Stop(testStopTimeout) }()
+
+	const stopBudget = testWaitDelay + 2*time.Second
+	select {
+	case err := <-stopReturned:
+		if err != nil {
+			t.Fatalf("Stop: %v", err)
+		}
+		t.Logf("Stop returned in %v", time.Since(stopStart))
+	case <-time.After(stopBudget):
+		t.Fatalf("Stop did not return within %v — cmd.Wait() likely hung on inherited pipe", stopBudget)
+	}
+
+	if got := p.State(); got != StateStopped {
+		t.Errorf("after Stop: expected state %s, got %s", StateStopped, got)
+	}
+
+	select {
+	case <-runErr:
+	case <-time.After(testReturnTimeout):
+		t.Errorf("Run did not return after Stop")
+	}
+}
+
+// TestProcessCommand_StopHonorsGracefulTimeout is a regression for the bug
+// where cmd.WaitDelay capped the graceful shutdown window. killProcess used to
+// cancel the cmd context to deliver SIGTERM, which starts cmd.WaitDelay
+// immediately; a process whose SIGTERM handler needs longer than WaitDelay to
+// finish was force-killed early even though Stop was given a much longer
+// timeout. The fix sends the signal directly so WaitDelay measures from process
+// exit (its inherited-pipe backstop role), leaving the graceful window to the
+// caller's Stop timeout.
+func TestProcessCommand_StopHonorsGracefulTimeout(t *testing.T) {
+	dir := t.TempDir()
+	marker := filepath.Join(dir, "graceful.done")
+	ready := filepath.Join(dir, "trap.ready")
+
+	// On SIGTERM, sleep past the (short) WaitDelay, then write the marker and
+	// exit cleanly. If WaitDelay still drove the kill, bash would be SIGKILLed
+	// mid-handler and the marker would never be written. The ready file is
+	// written only after the trap is installed so the test does not race
+	// SIGTERM ahead of it (CheckEndpoint:none marks ready before bash runs).
+	script := filepath.Join(dir, "graceful.sh")
+	body := fmt.Sprintf(
+		"#!/bin/bash\ncleanup() { sleep 0.6; echo done > %q; exit 0; }\ntrap cleanup SIGTERM\necho ready > %q\nwhile true; do sleep 0.1; done\n",
+		marker, ready,
+	)
+	if err := os.WriteFile(script, []byte(body), 0o755); err != nil {
+		t.Fatalf("WriteFile: %v", err)
+	}
+
+	p := newProcessCommand(t, config.ModelConfig{
+		Cmd:           script,
+		Proxy:         "http://127.0.0.1:1", // unused: health check disabled
+		CheckEndpoint: "none",
+	})
+	// WaitDelay shorter than the handler's 0.6s sleep, and far shorter than the
+	// Stop timeout below — this is the window the old code mis-killed in.
+	p.waitDelay = 200 * time.Millisecond
+
+	runErr := runAsync(t, p)
+
+	// Wait until the trap is installed before stopping.
+	trapDeadline := time.Now().Add(2 * time.Second)
+	for {
+		if _, err := os.Stat(ready); err == nil {
+			break
+		}
+		if time.Now().After(trapDeadline) {
+			t.Fatalf("script did not install SIGTERM trap in time")
+		}
+		time.Sleep(10 * time.Millisecond)
+	}
+
+	stopStart := time.Now()
+	if err := p.Stop(5 * time.Second); err != nil {
+		t.Fatalf("Stop: %v", err)
+	}
+	elapsed := time.Since(stopStart)
+
+	// The handler must have run to completion (marker written) rather than
+	// being force-killed at waitDelay.
+	if _, err := os.Stat(marker); err != nil {
+		t.Fatalf("graceful handler did not complete (marker missing): %v", err)
+	}
+	// And Stop must have waited for the handler (>~0.6s), not returned at the
+	// 200ms waitDelay.
+	if elapsed < 500*time.Millisecond {
+		t.Fatalf("Stop returned in %v — process was killed before its graceful handler finished", elapsed)
+	}
+
+	if got := p.State(); got != StateStopped {
+		t.Errorf("after Stop: expected state %s, got %s", StateStopped, got)
+	}
+	select {
+	case <-runErr:
+	case <-time.After(testReturnTimeout):
+		t.Errorf("Run did not return after Stop")
+	}
+}
+
+// TestProcessCommand_StopReapsForkedGrandchild verifies that stopping a forking
+// wrapper takes down the backgrounded grandchild too, rather than leaving it as
+// an orphan. The fix is Setpgid (runtime_unix.go): the wrapper leads its own
+// process group, so the stop signal is delivered to the whole group via the
+// negative PID and reaches the grandchild the wrapper never reaped.
+func TestProcessCommand_StopReapsForkedGrandchild(t *testing.T) {
+	skipIfNoSimpleResponder(t)
+
+	port := getFreePort(t)
+	dir := t.TempDir()
+	pidFile := filepath.Join(dir, "child.pid")
+
+	wrapper := filepath.Join(dir, "wrapper.sh")
+	script := fmt.Sprintf("#!/bin/bash\n%q -port %d -silent &\necho $! > %q\nwait\n",
+		simpleResponderPath, port, pidFile)
+	if err := os.WriteFile(wrapper, []byte(script), 0o755); err != nil {
+		t.Fatalf("WriteFile: %v", err)
+	}
+	t.Cleanup(func() { killChildFromPidFile(pidFile) })
+
+	p := newProcessCommand(t, config.ModelConfig{
+		Cmd:                wrapper,
+		Proxy:              fmt.Sprintf("http://127.0.0.1:%d", port),
+		CheckEndpoint:      "/health",
+		HealthCheckTimeout: 10,
+	})
+
+	runErr := runAsync(t, p)
+
+	// Read the grandchild PID the wrapper recorded.
+	var childPID int
+	deadline := time.Now().Add(2 * time.Second)
+	for {
+		data, err := os.ReadFile(pidFile)
+		if err == nil {
+			if pid, perr := strconv.Atoi(strings.TrimSpace(string(data))); perr == nil && pid > 0 {
+				childPID = pid
+				break
+			}
+		}
+		if time.Now().After(deadline) {
+			t.Fatalf("wrapper did not record grandchild PID")
+		}
+		time.Sleep(10 * time.Millisecond)
+	}
+
+	if err := p.Stop(testStopTimeout); err != nil {
+		t.Fatalf("Stop: %v", err)
+	}
+
+	// After Stop the grandchild must be gone. Signal 0 probes liveness without
+	// actually sending a signal; give it a brief window to exit after the
+	// group SIGTERM.
+	proc, err := os.FindProcess(childPID)
+	if err != nil {
+		t.Fatalf("FindProcess: %v", err)
+	}
+	gone := false
+	for i := 0; i < 100; i++ {
+		if err := proc.Signal(syscall.Signal(0)); err != nil {
+			gone = true
+			break
+		}
+		time.Sleep(10 * time.Millisecond)
+	}
+	if !gone {
+		t.Errorf("grandchild PID %d still alive after Stop — process group was not reaped", childPID)
+	}
+
+	select {
+	case <-runErr:
+	case <-time.After(testReturnTimeout):
+		t.Errorf("Run did not return after Stop")
+	}
+}
+
+// killChildFromPidFile reads a PID written by the wrapper script and SIGKILLs
+// it so leaked orphans don't accumulate between test runs. Best-effort.
+func killChildFromPidFile(pidFile string) {
+	data, err := os.ReadFile(pidFile)
+	if err != nil {
+		return
+	}
+	pid, err := strconv.Atoi(strings.TrimSpace(string(data)))
+	if err != nil || pid <= 0 {
+		return
+	}
+	proc, err := os.FindProcess(pid)
+	if err != nil {
+		return
+	}
+	_ = proc.Kill()
+}

internal/process/runtime_unix.go

@@ -4,9 +4,41 @@ package process
 
 import (
 	"os/exec"
+	"syscall"
 )
 
-// setProcAttributes sets platform-specific process attributes
+// setProcAttributes starts the upstream in its own process group (Setpgid) so
+// the entire process tree can be signalled at once via its negative PID. This
+// is what lets us reap a forked grandchild — e.g. a shell wrapper that
+// backgrounds the real binary and exits — instead of leaking it as an orphan
+// that holds the inherited stdout/stderr pipes open.
 func setProcAttributes(cmd *exec.Cmd) {
-	// No-op on Unix systems
+	cmd.SysProcAttr = &syscall.SysProcAttr{Setpgid: true}
+}
+
+// terminateProcessTree sends SIGTERM to the whole process group led by the
+// command, giving every process in the tree a chance to shut down gracefully.
+func terminateProcessTree(cmd *exec.Cmd) error {
+	return signalProcessTree(cmd, syscall.SIGTERM)
+}
+
+// killProcessTree sends SIGKILL to the whole process group, force-terminating
+// every process in the tree.
+func killProcessTree(cmd *exec.Cmd) error {
+	return signalProcessTree(cmd, syscall.SIGKILL)
+}
+
+// signalProcessTree signals the process group led by cmd.Process. Because the
+// child was started with Setpgid it is its own group leader (pgid == pid), so
+// targeting -pid reaches the child and every descendant still in the group.
+// Falls back to signalling just the child if the group send fails (e.g. the
+// group has already drained), so we never silently skip the signal.
+func signalProcessTree(cmd *exec.Cmd, sig syscall.Signal) error {
+	if cmd == nil || cmd.Process == nil {
+		return nil
+	}
+	if err := syscall.Kill(-cmd.Process.Pid, sig); err != nil {
+		return cmd.Process.Signal(sig)
+	}
+	return nil
 }

internal/process/runtime_windows.go

@@ -3,14 +3,51 @@
 package process
 
 import (
+	"fmt"
 	"os/exec"
 	"syscall"
 )
 
-// setProcAttributes sets platform-specific process attributes
+// setProcAttributes sets platform-specific process attributes. CREATE_NO_WINDOW
+// keeps the upstream from spawning its own console window.
 func setProcAttributes(cmd *exec.Cmd) {
 	cmd.SysProcAttr = &syscall.SysProcAttr{
 		HideWindow:    true,
 		CreationFlags: 0x08000000, // CREATE_NO_WINDOW
 	}
 }
+
+// terminateProcessTree requests a graceful shutdown of the whole process tree
+// rooted at cmd.Process. Windows has no SIGTERM or process-group signalling, so
+// we shell out to `taskkill /t`, which walks the child tree by PID — the
+// equivalent of signalling a Unix process group. Without /f, taskkill asks the
+// processes to close rather than force-killing them.
+func terminateProcessTree(cmd *exec.Cmd) error {
+	return taskkillProcessTree(cmd, false)
+}
+
+// killProcessTree force-terminates the whole process tree rooted at cmd.Process
+// via `taskkill /f /t`, so any descendant that ignored or outlived the graceful
+// request is killed alongside the parent rather than leaked as an orphan.
+func killProcessTree(cmd *exec.Cmd) error {
+	return taskkillProcessTree(cmd, true)
+}
+
+// taskkillProcessTree runs taskkill against cmd.Process.Pid. The /t flag
+// terminates the process together with any child processes it started, which is
+// the Windows analogue of signalling a Unix process group via its negative PID.
+// When force is true the /f flag force-kills; otherwise taskkill requests a
+// graceful close.
+func taskkillProcessTree(cmd *exec.Cmd, force bool) error {
+	if cmd == nil || cmd.Process == nil {
+		return nil
+	}
+	args := make([]string, 0, 4)
+	if force {
+		args = append(args, "/f")
+	}
+	args = append(args, "/t", "/pid", fmt.Sprintf("%d", cmd.Process.Pid))
+	kill := exec.Command("taskkill", args...)
+	setProcAttributes(kill)
+	return kill.Run()
+}

internal/process/treecleanup_other.go

@@ -0,0 +1,7 @@
+//go:build !windows
+
+package process
+
+// SetupTreeCleanup is a no-op on non-Windows platforms, where upstream process
+// teardown is handled via process-group signalling (see runtime_unix.go).
+func SetupTreeCleanup() error { return nil }

internal/process/treecleanup_windows.go

@@ -0,0 +1,50 @@
+//go:build windows
+
+package process
+
+import (
+	"fmt"
+	"unsafe"
+
+	"golang.org/x/sys/windows"
+)
+
+// SetupTreeCleanup assigns the current process to a Windows Job Object
+// configured with JOB_OBJECT_LIMIT_KILL_ON_JOB_CLOSE. Upstream processes
+// spawned afterwards are associated with the same job, so when llama-swap exits
+// for any reason — graceful shutdown, a forced second Ctrl+C, or a crash — the
+// OS terminates the whole job and reaps every child instead of leaving orphans
+// behind. It is the parent-side complement to the per-process teardown in
+// runtime_windows.go.
+//
+// The job handle is intentionally leaked for the lifetime of the process: the
+// kill-on-close behaviour fires when the last handle is released, which the OS
+// does when the process exits.
+func SetupTreeCleanup() error {
+	job, err := windows.CreateJobObject(nil, nil)
+	if err != nil {
+		return fmt.Errorf("CreateJobObject: %w", err)
+	}
+
+	info := windows.JOBOBJECT_EXTENDED_LIMIT_INFORMATION{
+		BasicLimitInformation: windows.JOBOBJECT_BASIC_LIMIT_INFORMATION{
+			LimitFlags: windows.JOB_OBJECT_LIMIT_KILL_ON_JOB_CLOSE,
+		},
+	}
+	if _, err := windows.SetInformationJobObject(
+		job,
+		windows.JobObjectExtendedLimitInformation,
+		uintptr(unsafe.Pointer(&info)),
+		uint32(unsafe.Sizeof(info)),
+	); err != nil {
+		windows.CloseHandle(job)
+		return fmt.Errorf("SetInformationJobObject: %w", err)
+	}
+
+	if err := windows.AssignProcessToJobObject(job, windows.CurrentProcess()); err != nil {
+		windows.CloseHandle(job)
+		return fmt.Errorf("AssignProcessToJobObject: %w", err)
+	}
+
+	return nil
+}

internal/router/base.go

@@ -11,6 +11,8 @@ import (
 	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
 	"github.com/mostlygeek/llama-swap/internal/process"
+	"github.com/mostlygeek/llama-swap/internal/router/scheduler"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
 type shutdownReq struct {
@@ -24,66 +26,38 @@ type unloadReq struct {
 	respond chan struct{}
 }
 
-type handlerReq struct {
-	model      string
-	ctx        context.Context
-	respond    chan handlerResp
-	positionCh chan int
-}
-
-type handlerResp struct {
-	handleFunc http.HandlerFunc
-	err        error
-}
-
-type swapDone struct {
-	modelID string
-	err     error
-}
-
-type serveDoneEvent struct {
-	modelID string
-}
-
-type activeSwap struct {
-	modelID string
-	evict   []string
-	waiters []handlerReq
-}
-
-// swapPlanner is the only piece of behaviour that differs between concrete
-// routers. baseRouter never inspects its internals.
-type swapPlanner interface {
-	// EvictionFor returns running model IDs that must be stopped before
-	// target can serve. alsoRunning lists models the baseRouter has already
-	// committed to loading (in-flight swaps) which the planner cannot see
-	// via process.State() yet. Pure decision; must not log.
-	EvictionFor(target string, alsoRunning []string) []string
-
-	// OnSwapStart runs once at the start of every swap. Planners may log
-	// their decision here at whatever verbosity they choose.
-	OnSwapStart(target string)
-}
-
-// baseRouter owns the channels, run-loop, and orchestration code shared by
-// every concrete router. Concrete routers embed *baseRouter and supply a
-// swapPlanner that captures how their eviction set is decided.
+// baseRouter owns the channels, run-loop, and process machinery shared by every
+// concrete router. Concrete routers embed *baseRouter and supply a
+// scheduler.Swapper describing how eviction sets are decided. baseRouter
+// implements scheduler.Effects so the scheduler can call back for side-effects.
 type baseRouter struct {
 	name      string
 	config    config.Config
 	processes map[string]process.Process
 	logger    *logmon.Monitor
-	planner   swapPlanner
+	schedule  scheduler.Scheduler
 
+	// shutdownCtx governs the request machinery: cancelling it tells grant()
+	// and ServeHTTP to stop granting and reject callers. It is deliberately
+	// separate from procCtx — see procCtx below.
 	shutdownCtx  context.Context
 	shutdownFn   context.CancelFunc
 	shuttingDown atomic.Bool
 
-	handlerCh   chan handlerReq
+	// procCtx is the parent context for every managed process and governs
+	// process lifetime only. handleShutdown stops processes gracefully via
+	// Stop() and cancels procCtx afterwards, so teardown is never a context
+	// cancel racing the graceful path (which collapsed the grace to 100ms and
+	// let the caller return before children were reaped — see process run loop).
+	procCtx    context.Context
+	procCancel context.CancelFunc
+
+	handlerCh   chan scheduler.HandlerReq
+	cancelCh    chan scheduler.HandlerReq
 	shutdownCh  chan shutdownReq
 	unloadCh    chan unloadReq
-	swapDoneCh  chan swapDone
-	serveDoneCh chan serveDoneEvent
+	swapDoneCh  chan scheduler.SwapDone
+	serveDoneCh chan scheduler.ServeDoneEvent
 
 	runDone chan struct{}
 
@@ -95,23 +69,38 @@ type baseRouter struct {
 	testProcessed chan struct{}
 }
 
-func newBaseRouter(name string, conf config.Config, processes map[string]process.Process, planner swapPlanner, logger *logmon.Monitor) *baseRouter {
+func newBaseRouter(
+	name string,
+	conf config.Config,
+	processes map[string]process.Process,
+	logger *logmon.Monitor,
+	planner scheduler.Swapper,
+) (*baseRouter, error) {
 	shutdownCtx, shutdownFn := context.WithCancel(context.Background())
-	return &baseRouter{
+	procCtx, procCancel := context.WithCancel(context.Background())
+	b := &baseRouter{
 		name:        name,
 		config:      conf,
 		processes:   processes,
 		logger:      logger,
-		planner:     planner,
 		shutdownCtx: shutdownCtx,
 		shutdownFn:  shutdownFn,
-		handlerCh:   make(chan handlerReq),
+		procCtx:     procCtx,
+		procCancel:  procCancel,
+		handlerCh:   make(chan scheduler.HandlerReq),
+		cancelCh:    make(chan scheduler.HandlerReq),
 		shutdownCh:  make(chan shutdownReq),
 		unloadCh:    make(chan unloadReq),
-		swapDoneCh:  make(chan swapDone),
-		serveDoneCh: make(chan serveDoneEvent),
+		swapDoneCh:  make(chan scheduler.SwapDone),
+		serveDoneCh: make(chan scheduler.ServeDoneEvent),
 		runDone:     make(chan struct{}),
 	}
+	sched, err := scheduler.New(conf, name, logger, planner, b)
+	if err != nil {
+		return nil, err
+	}
+	b.schedule = sched
+	return b, nil
 }
 
 func (b *baseRouter) notifyProcessed() {
@@ -123,30 +112,31 @@ func (b *baseRouter) notifyProcessed() {
 func (b *baseRouter) run() {
 	defer close(b.runDone)
 
-	active := make(map[string]*activeSwap)
-	inFlight := make(map[string]int)
-	var queued []handlerReq
-
 	for {
 		select {
 		case req := <-b.shutdownCh:
-			b.handleShutdown(req, active, queued)
+			b.handleShutdown(req)
 			return
 
 		case req := <-b.handlerCh:
-			b.handleRequest(req, active, inFlight, &queued)
+			b.schedule.OnRequest(req)
+			b.notifyProcessed()
+
+		case req := <-b.cancelCh:
+			b.schedule.OnCancel(req)
 			b.notifyProcessed()
 
 		case req := <-b.unloadCh:
-			b.handleUnload(req, active, inFlight, &queued)
+			b.schedule.OnUnload(req.targets, req.timeout)
+			close(req.respond)
 			b.notifyProcessed()
 
 		case ev := <-b.swapDoneCh:
-			b.handleSwapDone(ev, active, inFlight, &queued)
+			b.schedule.OnSwapDone(ev)
 			b.notifyProcessed()
 
 		case ev := <-b.serveDoneCh:
-			b.handleServeDone(ev, active, inFlight, &queued)
+			b.schedule.OnServeDone(ev)
 		}
 	}
 }
@@ -163,37 +153,68 @@ func (b *baseRouter) run() {
 // down, the send never lands, one of the other select cases fires, and we
 // report back that the grant did NOT happen.
 //
-// That distinction matters for in-flight bookkeeping — see grantHandler.
-func (b *baseRouter) grant(req handlerReq, resp handlerResp) bool {
+// That distinction matters for in-flight bookkeeping — see GrantServe.
+func (b *baseRouter) grant(req scheduler.HandlerReq, resp scheduler.HandlerResp) bool {
 	select {
-	case req.respond <- resp:
+	case req.Respond <- resp:
 		return true
-	case <-req.ctx.Done():
+	case <-req.Ctx.Done():
 		return false
 	case <-b.shutdownCtx.Done():
 		return false
 	}
 }
 
-// grantHandler is the "this caller can now use process p" path. It does
-// two things that must stay locked together:
-//
-//  1. Hand the caller a wrapped p.ServeHTTP (via trackedServe) so when the
-//     HTTP request finishes, the run loop hears about it.
-//  2. Bump inFlight[modelID] so the router knows this process is busy and
-//     refuses to evict it until the count comes back down.
-//
-// The increment is gated on grant() returning true. If grant() returns
-// false, the caller already walked away and trackedServe will never run —
-// which means no matching decrement will ever arrive on serveDoneCh.
-// Incrementing in that case would strand the counter at >0 forever and
-// the router would never again be willing to swap this model out.
-//
-// In short: increment if and only if we know a decrement is coming.
-func (b *baseRouter) grantHandler(req handlerReq, modelID string, p process.Process, inFlight map[string]int) {
-	if b.grant(req, handlerResp{handleFunc: b.trackedServe(modelID, p)}) {
-		inFlight[modelID]++
+// ModelState implements scheduler.Effects.
+func (b *baseRouter) ModelState(modelID string) (process.ProcessState, bool) {
+	p, ok := b.processes[modelID]
+	if !ok {
+		var zero process.ProcessState
+		return zero, false
 	}
+	return p.State(), true
+}
+
+// StartSwap implements scheduler.Effects, launching the swap goroutine.
+func (b *baseRouter) StartSwap(modelID string, evict []string) {
+	go b.doSwap(modelID, evict)
+}
+
+// GrantError implements scheduler.Effects.
+func (b *baseRouter) GrantError(req scheduler.HandlerReq, err error) {
+	b.grant(req, scheduler.HandlerResp{Err: err})
+}
+
+// GrantServe implements scheduler.Effects. It hands the caller a wrapped
+// p.ServeHTTP (via trackedServe) so the run loop hears about the request
+// finishing, and reports whether the caller received it. The scheduler bumps
+// its in-flight count only on a true return: if grant() returns false the
+// caller already walked away and trackedServe will never run, so no matching
+// decrement will ever arrive — incrementing would strand the counter at >0 and
+// the router would never again be willing to evict this model.
+func (b *baseRouter) GrantServe(req scheduler.HandlerReq, modelID string) bool {
+	p := b.processes[modelID]
+	return b.grant(req, scheduler.HandlerResp{HandleFunc: b.trackedServe(modelID, p)})
+}
+
+// StopProcesses implements scheduler.Effects, stopping the named processes in
+// parallel and blocking until all have stopped.
+func (b *baseRouter) StopProcesses(timeout time.Duration, ids []string) {
+	var wg sync.WaitGroup
+	for _, id := range ids {
+		p, ok := b.processes[id]
+		if !ok {
+			continue
+		}
+		wg.Add(1)
+		go func(id string, p process.Process) {
+			defer wg.Done()
+			if err := p.Stop(timeout); err != nil {
+				b.logger.Warnf("%s: stopping %s failed: %v", b.name, id, err)
+			}
+		}(id, p)
+	}
+	wg.Wait()
 }
 
 // trackedServe is the wrapper that closes the loop on in-flight tracking.
@@ -210,7 +231,7 @@ func (b *baseRouter) trackedServe(modelID string, p process.Process) http.Handle
 	return func(w http.ResponseWriter, r *http.Request) {
 		defer func() {
 			select {
-			case b.serveDoneCh <- serveDoneEvent{modelID: modelID}:
+			case b.serveDoneCh <- scheduler.ServeDoneEvent{ModelID: modelID}:
 			case <-b.shutdownCtx.Done():
 			}
 		}()
@@ -218,240 +239,6 @@ func (b *baseRouter) trackedServe(modelID string, p process.Process) http.Handle
 	}
 }
 
-// handleRequest decides what to do with one incoming ServeHTTP request. It is
-// called from run() and never blocks indefinitely: any work that has to wait
-// (starting a process, stopping siblings, waiting for ready) is deferred to
-// a swap goroutine and reported back via swapDoneCh.
-//
-// The decision tree, in order:
-//
-//  1. Unknown model — respond with ErrNoLocalModelFound and move on.
-//  2. A swap to the same model is already in flight — attach this waiter so
-//     one swap serves all callers that asked for the same model.
-//  3. Fast path — the target process is already ready, the planner sees
-//     nothing to evict, and no in-flight swap is evicting it. Hand back its
-//     ServeHTTP immediately (wrapped so the run loop knows when it ends).
-//  4. Would collide with an in-flight swap (we'd stop their target, or
-//     they're stopping us) — park in the queue for handleSwapDone to drain.
-//  5. Would evict a process that is still handling requests — park in the
-//     queue. handleServeDone will retry when the busy process drains.
-//  6. Otherwise — start a new swap. This may run in parallel with other
-//     active swaps when their evict sets don't intersect.
-func (b *baseRouter) handleRequest(req handlerReq, active map[string]*activeSwap, inFlight map[string]int, queued *[]handlerReq) {
-	// (1) Unknown model.
-	p, ok := b.processes[req.model]
-	if !ok {
-		b.logger.Debugf("%s: model %s not handled by this router", b.name, req.model)
-		b.grant(req, handlerResp{err: ErrNoLocalModelFound})
-		return
-	}
-
-	// (2) Join an in-flight swap for the same model.
-	if s, ok := active[req.model]; ok {
-		b.logger.Debugf("%s: joining in-flight swap for model %s (%d waiters)", b.name, req.model, len(s.waiters)+1)
-		s.waiters = append(s.waiters, req)
-		return
-	}
-
-	evict := b.planner.EvictionFor(req.model, activeTargets(active, req.model))
-
-	// (3) Fast path: ready, nothing to evict, and nobody is evicting us.
-	if p.State() == process.StateReady && len(evict) == 0 && !collidesWith(req.model, evict, active) {
-		b.logger.Debugf("%s: fast-path serving model %s (already ready)", b.name, req.model)
-		b.grantHandler(req, req.model, p, inFlight)
-		return
-	}
-
-	// (4) Collision with an in-flight swap — queue.
-	if collidesWith(req.model, evict, active) {
-		b.logger.Debugf("%s: queuing request for model %s (collides with in-flight swap)", b.name, req.model)
-		*queued = append(*queued, req)
-		b.broadcastQueuePositions(*queued)
-		return
-	}
-
-	// (5) Would evict a busy process — queue until it drains.
-	if conflictsWithInFlight(evict, inFlight) {
-		b.logger.Debugf("%s: queuing request for model %s (would evict in-flight process)", b.name, req.model)
-		*queued = append(*queued, req)
-		b.broadcastQueuePositions(*queued)
-		return
-	}
-
-	// (6) Start a new (possibly parallel) swap.
-	b.logger.Debugf("%s: starting swap for model %s, evicting %v", b.name, req.model, evict)
-	s := b.startSwap(req, evict)
-	active[s.modelID] = s
-}
-
-// handleSwapDone is called from run() when a swap goroutine reports that it
-// has finished. It fans out the result to every waiter that joined this swap,
-// removes the swap from the active map, and then walks the queue once,
-// promoting any items that no longer collide with the remaining active set.
-// FIFO order is preserved: items still blocked stay in place.
-func (b *baseRouter) handleSwapDone(ev swapDone, active map[string]*activeSwap, inFlight map[string]int, queued *[]handlerReq) {
-	s, ok := active[ev.modelID]
-	if !ok {
-		return
-	}
-	delete(active, ev.modelID)
-
-	for _, w := range s.waiters {
-		if ev.err != nil {
-			b.grant(w, handlerResp{err: ev.err})
-		} else {
-			p := b.processes[ev.modelID]
-			b.grantHandler(w, ev.modelID, p, inFlight)
-		}
-	}
-
-	b.drainQueue(active, inFlight, queued)
-}
-
-// handleServeDone is called from run() each time a tracked ServeHTTP
-// finishes. It decrements the per-model in-flight count and, when that
-// drops to zero, retries the queue: requests whose swap was deferred
-// because they would have evicted this (now-idle) process can now proceed.
-func (b *baseRouter) handleServeDone(ev serveDoneEvent, active map[string]*activeSwap, inFlight map[string]int, queued *[]handlerReq) {
-	inFlight[ev.modelID]--
-	if inFlight[ev.modelID] <= 0 {
-		delete(inFlight, ev.modelID)
-		b.drainQueue(active, inFlight, queued)
-	}
-}
-
-// drainQueue walks the queued requests in order, re-running the handleRequest
-// decision tree against the (now smaller) active set. Items that can now start
-// or join become satisfied; items still blocked remain queued in original
-// order so they get another chance on the next swap completion.
-func (b *baseRouter) drainQueue(active map[string]*activeSwap, inFlight map[string]int, queued *[]handlerReq) {
-	if len(*queued) == 0 {
-		return
-	}
-	pending := *queued
-	var remaining []handlerReq
-	for _, req := range pending {
-		p, ok := b.processes[req.model]
-		if !ok {
-			b.grant(req, handlerResp{err: ErrNoLocalModelFound})
-			continue
-		}
-		if s, ok := active[req.model]; ok {
-			b.logger.Debugf("%s: queued request for model %s now joining in-flight swap", b.name, req.model)
-			s.waiters = append(s.waiters, req)
-			continue
-		}
-		evict := b.planner.EvictionFor(req.model, activeTargets(active, req.model))
-		if p.State() == process.StateReady && len(evict) == 0 && !collidesWith(req.model, evict, active) {
-			b.logger.Debugf("%s: queued request for model %s now served fast-path", b.name, req.model)
-			b.grantHandler(req, req.model, p, inFlight)
-			continue
-		}
-		if collidesWith(req.model, evict, active) {
-			remaining = append(remaining, req)
-			continue
-		}
-		if conflictsWithInFlight(evict, inFlight) {
-			remaining = append(remaining, req)
-			continue
-		}
-		b.logger.Debugf("%s: queued request for model %s now starting swap, evicting %v", b.name, req.model, evict)
-		s := b.startSwap(req, evict)
-		active[s.modelID] = s
-	}
-	*queued = remaining
-	b.broadcastQueuePositions(*queued)
-}
-
-// broadcastQueuePositions sends each queued request its current 1-indexed
-// position. Sends are non-blocking: if the channel is full, the old value is
-// drained first so the consumer always sees the latest position.
-func (b *baseRouter) broadcastQueuePositions(queued []handlerReq) {
-	for i, req := range queued {
-		pos := i + 1
-		select {
-		case req.positionCh <- pos:
-		default:
-			select {
-			case <-req.positionCh:
-			default:
-			}
-			select {
-			case req.positionCh <- pos:
-			default:
-			}
-		}
-	}
-}
-
-func (b *baseRouter) startSwap(initial handlerReq, evict []string) *activeSwap {
-	swap := &activeSwap{
-		modelID: initial.model,
-		evict:   evict,
-		waiters: []handlerReq{initial},
-	}
-	b.planner.OnSwapStart(initial.model)
-	go b.doSwap(initial.model, evict)
-	return swap
-}
-
-// activeTargets returns the IDs of every in-flight swap target except exclude.
-// baseRouter passes this to the planner so eviction decisions account for
-// models that have been committed to but have not yet transitioned to
-// StateStarting in their process state machine.
-func activeTargets(active map[string]*activeSwap, exclude string) []string {
-	if len(active) == 0 {
-		return nil
-	}
-	out := make([]string, 0, len(active))
-	for id := range active {
-		if id == exclude {
-			continue
-		}
-		out = append(out, id)
-	}
-	return out
-}
-
-// collidesWith reports whether a new swap with this target and evict set can
-// safely run alongside the currently active swaps. Same-target callers should
-// JOIN (handled before this) — they do not collide with themselves.
-func collidesWith(target string, evict []string, active map[string]*activeSwap) bool {
-	for id, s := range active {
-		if id == target {
-			continue
-		}
-		if containsString(evict, id) {
-			return true
-		}
-		if containsString(s.evict, target) {
-			return true
-		}
-	}
-	return false
-}
-
-// conflictsWithInFlight reports whether any model in evict is still handling
-// requests. Stopping a busy process would cancel its callers' connections,
-// so the router defers the swap until those callers finish.
-func conflictsWithInFlight(evict []string, inFlight map[string]int) bool {
-	for _, m := range evict {
-		if inFlight[m] > 0 {
-			return true
-		}
-	}
-	return false
-}
-
-func containsString(xs []string, s string) bool {
-	for _, x := range xs {
-		if x == s {
-			return true
-		}
-	}
-	return false
-}
-
 func (b *baseRouter) doSwap(modelID string, toStop []string) {
 	timeout := b.healthCheckTimeout()
 
@@ -479,29 +266,24 @@ func (b *baseRouter) doSwap(modelID string, toStop []string) {
 	err := target.WaitReady(b.shutdownCtx)
 
 	select {
-	case b.swapDoneCh <- swapDone{modelID: modelID, err: err}:
+	case b.swapDoneCh <- scheduler.SwapDone{ModelID: modelID, Err: err}:
 	case <-b.shutdownCtx.Done():
 	}
 }
 
-func (b *baseRouter) handleShutdown(req shutdownReq, active map[string]*activeSwap, queued []handlerReq) {
+func (b *baseRouter) handleShutdown(req shutdownReq) {
 	shutdownErr := fmt.Errorf("%s is shutting down", b.name)
 
 	// Cancel shutdownCtx first so any waiter that is currently parked on
 	// its respond channel can exit via its own shutdownCtx.Done() branch.
-	// The grant calls below then either land (waiter happened to receive
+	// The OnShutdown grants below then either land (waiter happened to receive
 	// before noticing shutdown) or fall through immediately via grant's
 	// shutdownCtx case — either way the waiter sees a non-OK response.
+	// This does NOT touch processes: their lifetime is procCtx, cancelled
+	// only after the graceful Stop() calls below have reaped them.
 	b.shutdownFn()
 
-	for _, s := range active {
-		for _, w := range s.waiters {
-			b.grant(w, handlerResp{err: shutdownErr})
-		}
-	}
-	for _, w := range queued {
-		b.grant(w, handlerResp{err: shutdownErr})
-	}
+	b.schedule.OnShutdown(shutdownErr)
 
 	stopTimeout := req.timeout
 	if stopTimeout <= 0 {
@@ -535,6 +317,11 @@ func (b *baseRouter) handleShutdown(req shutdownReq, active map[string]*activeSw
 		<-done
 	}
 
+	// Every process is stopped (children reaped via Stop()). Cancel procCtx so
+	// the process run-loop goroutines exit; they are already StateStopped, so
+	// this is a clean no-op kill rather than a forced teardown.
+	b.procCancel()
+
 	req.respond <- nil
 }
 
@@ -607,75 +394,6 @@ func (b *baseRouter) Unload(timeout time.Duration, models ...string) {
 	<-req.respond
 }
 
-// handleUnload runs on the run loop in response to an Unload call. It
-// reconciles router-owned state with the impending Stop, then performs
-// the Stop synchronously so callers of Unload remain blocked until each
-// targeted process has actually exited.
-func (b *baseRouter) handleUnload(req unloadReq, active map[string]*activeSwap, inFlight map[string]int, queued *[]handlerReq) {
-	unloadErr := fmt.Errorf("%s: model unloaded", b.name)
-
-	targetSet := make(map[string]bool, len(req.targets))
-	for _, id := range req.targets {
-		targetSet[id] = true
-	}
-
-	// Release waiters of any in-flight swap whose target is being
-	// unloaded. The swap goroutine itself is left to finish on its own;
-	// when its swapDone arrives, handleSwapDone will find no entry in
-	// active and silently drop it.
-	for id := range targetSet {
-		s, ok := active[id]
-		if !ok {
-			continue
-		}
-		for _, w := range s.waiters {
-			b.grant(w, handlerResp{err: unloadErr})
-		}
-		delete(active, id)
-	}
-
-	// Drop queued requests addressed to unloaded models. Requests for
-	// other models stay queued and may benefit from drainQueue at the end.
-	if len(*queued) > 0 {
-		kept := (*queued)[:0]
-		for _, w := range *queued {
-			if targetSet[w.model] {
-				b.grant(w, handlerResp{err: unloadErr})
-				continue
-			}
-			kept = append(kept, w)
-		}
-		*queued = kept
-	}
-
-	// Stop the targeted processes. Done synchronously so Unload's caller
-	// can rely on "after Unload returns, the process is stopped". inFlight
-	// is intentionally NOT cleared here: each dying handler will fire its
-	// trackedServe defer and reach handleServeDone in the normal way once
-	// the run loop is free again.
-	var wg sync.WaitGroup
-	for id := range targetSet {
-		p, ok := b.processes[id]
-		if !ok {
-			continue
-		}
-		wg.Add(1)
-		go func(id string, p process.Process) {
-			defer wg.Done()
-			if err := p.Stop(req.timeout); err != nil {
-				b.logger.Warnf("%s: unloading %s failed: %v", b.name, id, err)
-			}
-		}(id, p)
-	}
-	wg.Wait()
-
-	// Removing entries from active above may have unblocked queued
-	// requests that previously collided with the now-cancelled swaps.
-	b.drainQueue(active, inFlight, queued)
-
-	close(req.respond)
-}
-
 func (b *baseRouter) Shutdown(timeout time.Duration) error {
 	if !b.shuttingDown.CompareAndSwap(false, true) {
 		return fmt.Errorf("%s shutdown already in progress", b.name)
@@ -691,24 +409,24 @@ func (b *baseRouter) Shutdown(timeout time.Duration) error {
 
 func (b *baseRouter) ServeHTTP(w http.ResponseWriter, req *http.Request) {
 	if b.shuttingDown.Load() {
-		SendError(w, req, fmt.Errorf("%s is shutting down", b.name))
+		shared.SendError(w, req, fmt.Errorf("%s is shutting down", b.name))
 		return
 	}
 
-	data, err := FetchContext(req, b.config)
+	data, err := shared.FetchContext(req, b.config)
 	if err != nil {
-		SendError(w, req, err)
+		shared.SendError(w, req, err)
 		return
 	}
 
-	hr := handlerReq{
-		model: data.ModelID,
-		ctx:   req.Context(),
-		// Unbuffered: a successful send on respond proves the waiter is
+	hr := scheduler.HandlerReq{
+		Model: data.ModelID,
+		Ctx:   req.Context(),
+		// Unbuffered: a successful send on Respond proves the waiter is
 		// alive and consuming. grant() relies on this to avoid handing a
 		// handleFunc to a cancelled waiter and leaking the inFlight count.
-		respond:    make(chan handlerResp),
-		positionCh: make(chan int, 1),
+		Respond:    make(chan scheduler.HandlerResp),
+		PositionCh: make(chan int, 1),
 	}
 
 	select {
@@ -716,7 +434,7 @@ func (b *baseRouter) ServeHTTP(w http.ResponseWriter, req *http.Request) {
 	case <-req.Context().Done():
 		return
 	case <-b.shutdownCtx.Done():
-		SendError(w, req, fmt.Errorf("%s is shutting down", b.name))
+		shared.SendError(w, req, fmt.Errorf("%s is shutting down", b.name))
 		return
 	}
 
@@ -736,7 +454,7 @@ func (b *baseRouter) ServeHTTP(w http.ResponseWriter, req *http.Request) {
 		go func() {
 			for {
 				select {
-				case pos := <-hr.positionCh:
+				case pos := <-hr.PositionCh:
 					lw.setUpdate(fmt.Sprintf("Queue position: #%d", pos))
 				case <-swapCtx.Done():
 					return
@@ -745,31 +463,43 @@ func (b *baseRouter) ServeHTTP(w http.ResponseWriter, req *http.Request) {
 		}()
 	}
 
-	var resp handlerResp
-	select {
-	case resp = <-hr.respond:
+	// finishLoading stops the loading stream and fences its goroutine off from
+	// the ResponseWriter before the real handler (or ServeHTTP's return)
+	// reclaims it. release() must run even when waitForCompletion times out:
+	// otherwise a still-streaming goroutine flushes a finalized response and
+	// panics on the recycled *bufio.Writer.
+	finishLoading := func() {
 		cancelLoad()
 		if lw != nil {
 			lw.waitForCompletion(1 * time.Second)
+			lw.release()
 		}
+	}
+
+	var resp scheduler.HandlerResp
+	select {
+	case resp = <-hr.Respond:
+		finishLoading()
 	case <-req.Context().Done():
-		cancelLoad()
-		if lw != nil {
-			lw.waitForCompletion(1 * time.Second)
+		finishLoading()
+		// Notify the scheduler so it can prune this request from its queue
+		// and swap waiters. Without this, a queued request whose client left
+		// would sit in the scheduler until drainQueue eventually starts a
+		// wasted model load for it.
+		select {
+		case b.cancelCh <- hr:
+		case <-b.shutdownCtx.Done():
 		}
 		return
 	case <-b.shutdownCtx.Done():
-		cancelLoad()
-		if lw != nil {
-			lw.waitForCompletion(1 * time.Second)
-		}
-		SendError(w, req, fmt.Errorf("%s is shutting down", b.name))
+		finishLoading()
+		shared.SendError(w, req, fmt.Errorf("%s is shutting down", b.name))
 		return
 	}
 
-	if resp.err != nil {
-		SendError(w, req, resp.err)
+	if resp.Err != nil {
+		shared.SendError(w, req, resp.Err)
 		return
 	}
-	resp.handleFunc(w, req)
+	resp.HandleFunc(w, req)
 }

internal/router/base_test.go

@@ -5,35 +5,34 @@ import (
 	"io"
 	"net/http"
 	"net/http/httptest"
-	"sync"
 	"testing"
 	"time"
 
 	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
 	"github.com/mostlygeek/llama-swap/internal/process"
+	"github.com/mostlygeek/llama-swap/internal/router/scheduler"
 )
 
-// stubPlanner is a swapPlanner that returns a fixed eviction list per target
-// and never logs. It lets the base-router tests cover shared run-loop
-// behaviour without dragging in either real router's eviction rules.
-type stubPlanner struct {
-	evict map[string][]string
-}
+// These tests cover baseRouter's own machinery — the run loop, process
+// lifecycle (doSwap), grant/ServeHTTP plumbing, Unload, and Shutdown. The
+// scheduling decision logic (queueing, collation, eviction collisions) lives in
+// the scheduler package and is tested directly there; see fifo_test.go.
 
-func (s *stubPlanner) EvictionFor(target string, _ []string) []string {
-	if s.evict == nil {
-		return nil
-	}
-	return s.evict[target]
-}
+// stubPlanner evicts nothing. baseRouter tests drive the run loop through the
+// default FIFO scheduler without exercising any particular eviction policy.
+type stubPlanner struct{}
 
-func (s *stubPlanner) OnSwapStart(string) {}
+func (s *stubPlanner) EvictionFor(string, []string) []string { return nil }
+func (s *stubPlanner) OnSwapStart(string, []string)          {}
 
-func newTestBase(t *testing.T, processes map[string]process.Process, planner swapPlanner) *baseRouter {
+func newTestBase(t *testing.T, processes map[string]process.Process, planner scheduler.Swapper) *baseRouter {
 	t.Helper()
 	conf := config.Config{HealthCheckTimeout: 5}
-	b := newBaseRouter("test", conf, processes, planner, logmon.NewWriter(io.Discard))
+	b, err := newBaseRouter("test", conf, processes, logmon.NewWriter(io.Discard), planner)
+	if err != nil {
+		t.Fatalf("newBaseRouter: %v", err)
+	}
 	b.testProcessed = make(chan struct{}, 64)
 	go b.run()
 	t.Cleanup(func() {
@@ -157,114 +156,6 @@ func TestBaseRouter_Unload_StopsInParallel(t *testing.T) {
 	}
 }
 
-// TestBaseRouter_Unload_ReleasesActiveSwapWaiters verifies that Unload
-// rejoins router state: a request whose swap to the unloaded model is
-// still in progress receives an error, instead of being abandoned
-// against a process that's about to vanish.
-func TestBaseRouter_Unload_ReleasesActiveSwapWaiters(t *testing.T) {
-	a := newFakeProcess("a")
-	// autoReady=false: the swap parks on WaitReady so we can interrupt
-	// it with Unload before it completes.
-
-	b := newTestBase(t, map[string]process.Process{"a": a}, &stubPlanner{})
-
-	w := httptest.NewRecorder()
-	done := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w, newRequest("a"))
-		close(done)
-	}()
-	waitProcessed(t, b.testProcessed, 1) // handlerReq absorbed; swap started
-	<-a.runStarted
-
-	b.Unload(time.Second, "a")
-
-	select {
-	case <-done:
-	case <-time.After(2 * time.Second):
-		t.Fatal("ServeHTTP did not return after Unload")
-	}
-	if w.Code == http.StatusOK {
-		t.Errorf("expected non-OK status after Unload, got %d body=%q", w.Code, w.Body.String())
-	}
-	if a.State() != process.StateStopped {
-		t.Errorf("a state=%q want stopped", a.State())
-	}
-}
-
-// TestBaseRouter_Unload_DropsQueuedRequests verifies that queued requests
-// for an unloaded model receive an error rather than sitting forever in
-// the queue against state the router no longer maintains.
-func TestBaseRouter_Unload_DropsQueuedRequests(t *testing.T) {
-	a := newFakeProcess("a")
-	pb := newFakeProcess("b")
-	// Loading B evicts A — so a request for B while A is loading queues.
-	planner := &stubPlanner{evict: map[string][]string{"b": {"a"}}}
-	b := newTestBase(t, map[string]process.Process{"a": a, "b": pb}, planner)
-
-	// r1 starts the swap to A and parks on WaitReady (autoReady=false).
-	w1 := httptest.NewRecorder()
-	done1 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w1, newRequest("a"))
-		close(done1)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-	<-a.runStarted
-
-	// r2 for B collides with A's in-flight swap and queues.
-	w2 := httptest.NewRecorder()
-	done2 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w2, newRequest("b"))
-		close(done2)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-
-	// Unload B — r2 (queued, targeting B) must be released with an error.
-	b.Unload(time.Second, "b")
-
-	select {
-	case <-done2:
-	case <-time.After(2 * time.Second):
-		t.Fatal("queued B request did not return after Unload(b)")
-	}
-	if w2.Code == http.StatusOK {
-		t.Errorf("queued B request: expected non-OK status, got %d", w2.Code)
-	}
-	if got := pb.runCalls.Load(); got != 0 {
-		t.Errorf("b.runCalls=%d want 0 (B should never have been started)", got)
-	}
-
-	// Release r1 so the test cleans up cleanly.
-	a.markReady()
-	select {
-	case <-done1:
-	case <-time.After(2 * time.Second):
-		t.Fatal("r1 did not complete after a.markReady")
-	}
-}
-
-func TestBaseRouter_FastPath(t *testing.T) {
-	a := newFakeProcess("a")
-	a.markReady()
-
-	b := newTestBase(t, map[string]process.Process{"a": a}, &stubPlanner{})
-
-	w := httptest.NewRecorder()
-	b.ServeHTTP(w, newRequest("a"))
-
-	if w.Code != http.StatusOK {
-		t.Fatalf("status=%d body=%q", w.Code, w.Body.String())
-	}
-	if got := a.serveCalls.Load(); got != 1 {
-		t.Errorf("serveCalls=%d want 1", got)
-	}
-	if got := a.runCalls.Load(); got != 0 {
-		t.Errorf("runCalls=%d want 0 (fast path should not start)", got)
-	}
-}
-
 func TestBaseRouter_OnDemandStart(t *testing.T) {
 	a := newFakeProcess("a")
 	a.autoReady = true
@@ -285,43 +176,6 @@ func TestBaseRouter_OnDemandStart(t *testing.T) {
 	}
 }
 
-func TestBaseRouter_ConcurrentSameModel(t *testing.T) {
-	a := newFakeProcess("a")
-	// autoReady=false so the swap parks on WaitReady until we release it.
-
-	b := newTestBase(t, map[string]process.Process{"a": a}, &stubPlanner{})
-
-	const N = 5
-	var wg sync.WaitGroup
-	codes := make([]int, N)
-	for i := 0; i < N; i++ {
-		wg.Add(1)
-		go func(i int) {
-			defer wg.Done()
-			w := httptest.NewRecorder()
-			b.ServeHTTP(w, newRequest("a"))
-			codes[i] = w.Code
-		}(i)
-	}
-
-	waitProcessed(t, b.testProcessed, N) // all N handlerReqs absorbed by run()
-	<-a.runStarted                       // swap goroutine reached Run()
-	a.markReady()
-	wg.Wait()
-
-	for i, c := range codes {
-		if c != http.StatusOK {
-			t.Errorf("request %d: status=%d", i, c)
-		}
-	}
-	if got := a.runCalls.Load(); got != 1 {
-		t.Errorf("runCalls=%d want 1 (single swap should issue one Run)", got)
-	}
-	if got := a.serveCalls.Load(); got != N {
-		t.Errorf("serveCalls=%d want %d", got, N)
-	}
-}
-
 func TestBaseRouter_ContextCancel(t *testing.T) {
 	a := newFakeProcess("a")
 	// autoReady=false so swap parks forever until we mark ready.
@@ -364,459 +218,6 @@ func TestBaseRouter_ContextCancel(t *testing.T) {
 	}
 }
 
-func TestBaseRouter_QueuedDifferentModel(t *testing.T) {
-	a := newFakeProcess("a")
-	pa := newFakeProcess("b")
-
-	// Loading b must stop a.
-	planner := &stubPlanner{evict: map[string][]string{"b": {"a"}}}
-	b := newTestBase(t, map[string]process.Process{"a": a, "b": pa}, planner)
-
-	// First request starts a swap to A; A's autoReady=false so it parks.
-	w1 := httptest.NewRecorder()
-	done1 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w1, newRequest("a"))
-		close(done1)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-	<-a.runStarted
-
-	// Second request for B should queue while A's swap is in flight.
-	w2 := httptest.NewRecorder()
-	done2 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w2, newRequest("b"))
-		close(done2)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-
-	if got := pa.runCalls.Load(); got != 0 {
-		t.Errorf("b started early: runCalls=%d want 0 while A's swap is pending", got)
-	}
-
-	// Release A's swap. B's swap should then run.
-	a.markReady()
-	waitProcessed(t, b.testProcessed, 1) // swapDone for A → B's swap kicked off
-	<-pa.runStarted
-
-	select {
-	case <-done1:
-	case <-time.After(time.Second):
-		t.Fatal("A request did not complete")
-	}
-	pa.markReady()
-	select {
-	case <-done2:
-	case <-time.After(time.Second):
-		t.Fatal("queued B request did not complete after A's swap")
-	}
-	if w2.Code != http.StatusOK {
-		t.Errorf("B status=%d body=%q", w2.Code, w2.Body.String())
-	}
-	if got := a.stopCalls.Load(); got != 1 {
-		t.Errorf("a.stopCalls=%d want 1 (B's swap must stop A)", got)
-	}
-}
-
-// TestBaseRouter_QueueCollation verifies that incoming requests of the form
-// a, b, c, a, b, c collapse into three swaps (one per model) and that the
-// second request for each model rides the fast path — either joining the
-// active swap, or being pulled out of the queue when handleSwapDone promotes
-// the next model.
-func TestBaseRouter_QueueCollation(t *testing.T) {
-	a := newFakeProcess("a")
-	pb := newFakeProcess("b")
-	pc := newFakeProcess("c")
-
-	// Each model evicts the other two so all swaps are mutually exclusive.
-	planner := &stubPlanner{evict: map[string][]string{
-		"a": {"b", "c"},
-		"b": {"a", "c"},
-		"c": {"a", "b"},
-	}}
-	b := newTestBase(t, map[string]process.Process{"a": a, "b": pb, "c": pc}, planner)
-
-	var (
-		completedMu sync.Mutex
-		completed   []string
-	)
-	record := func(id string) {
-		completedMu.Lock()
-		defer completedMu.Unlock()
-		completed = append(completed, id)
-	}
-
-	ids := []string{"a", "b", "c", "a", "b", "c"}
-	var wg sync.WaitGroup
-	for _, id := range ids {
-		id := id
-		wg.Add(1)
-		go func() {
-			defer wg.Done()
-			w := httptest.NewRecorder()
-			b.ServeHTTP(w, newRequest(id))
-			if w.Code != http.StatusOK {
-				t.Errorf("%s: status=%d body=%q", id, w.Code, w.Body.String())
-				return
-			}
-			record(id)
-		}()
-		// Wait for run() to absorb this request before launching the next,
-		// so handlerCh receives them in launch order.
-		waitProcessed(t, b.testProcessed, 1)
-	}
-
-	// All 6 are now parked in run()'s waiters/queue. Release each swap in
-	// sequence, waiting deterministically for each promotion to fire.
-	<-a.runStarted
-	a.markReady()
-	waitProcessed(t, b.testProcessed, 1) // swapDone(a) → b swap kicked off
-
-	<-pb.runStarted
-	pb.markReady()
-	waitProcessed(t, b.testProcessed, 1) // swapDone(b) → c swap kicked off
-
-	<-pc.runStarted
-	pc.markReady()
-	wg.Wait()
-
-	if got := len(completed); got != 6 {
-		t.Fatalf("completed=%v want 6", completed)
-	}
-
-	// run() fans out responses in model-grouped order (a1,a2 → b1,b2 → c1,c2)
-	// but waiter goroutines may be scheduled in any order after their respond
-	// channel fires, so completion order isn't deterministic. Per-model counts
-	// (combined with the runCalls checks below) are sufficient to prove queue
-	// collation collapsed each pair into a single swap.
-	aDone, bDone, cDone := 0, 0, 0
-	for _, id := range completed {
-		switch id {
-		case "a":
-			aDone++
-		case "b":
-			bDone++
-		case "c":
-			cDone++
-		}
-	}
-	if aDone != 2 || bDone != 2 || cDone != 2 {
-		t.Errorf("per-model counts: a=%d b=%d c=%d, want 2 each (order=%v)", aDone, bDone, cDone, completed)
-	}
-
-	// Single swap per model — the second request for each must have ridden
-	// the fast path (joined active swap or joined a queued sibling), not
-	// triggered an extra Run.
-	if got := a.runCalls.Load(); got != 1 {
-		t.Errorf("a.runCalls=%d want 1", got)
-	}
-	if got := pb.runCalls.Load(); got != 1 {
-		t.Errorf("b.runCalls=%d want 1", got)
-	}
-	if got := pc.runCalls.Load(); got != 1 {
-		t.Errorf("c.runCalls=%d want 1", got)
-	}
-}
-
-// TestBaseRouter_ConcurrentDisjointSwaps verifies that two requests with
-// non-conflicting evict sets are loaded in parallel: both Run() calls happen
-// before either process is marked ready.
-func TestBaseRouter_ConcurrentDisjointSwaps(t *testing.T) {
-	a := newFakeProcess("a")
-	pb := newFakeProcess("b")
-
-	// Empty evict sets for both: they can load in parallel.
-	b := newTestBase(t, map[string]process.Process{"a": a, "b": pb}, &stubPlanner{})
-
-	w1 := httptest.NewRecorder()
-	done1 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w1, newRequest("a"))
-		close(done1)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-
-	w2 := httptest.NewRecorder()
-	done2 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w2, newRequest("b"))
-		close(done2)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-
-	// Both swaps must have reached Run() before either is marked ready —
-	// proves they ran in parallel rather than serializing.
-	<-a.runStarted
-	<-pb.runStarted
-
-	a.markReady()
-	pb.markReady()
-
-	select {
-	case <-done1:
-	case <-time.After(time.Second):
-		t.Fatal("request A did not complete")
-	}
-	select {
-	case <-done2:
-	case <-time.After(time.Second):
-		t.Fatal("request B did not complete")
-	}
-
-	if w1.Code != http.StatusOK {
-		t.Errorf("A status=%d body=%q", w1.Code, w1.Body.String())
-	}
-	if w2.Code != http.StatusOK {
-		t.Errorf("B status=%d body=%q", w2.Code, w2.Body.String())
-	}
-	if got := a.stopCalls.Load(); got != 0 {
-		t.Errorf("a.stopCalls=%d want 0 (parallel swap, no eviction)", got)
-	}
-	if got := pb.stopCalls.Load(); got != 0 {
-		t.Errorf("b.stopCalls=%d want 0 (parallel swap, no eviction)", got)
-	}
-}
-
-// TestBaseRouter_QueueDrainPromotesMultiple verifies that completing one swap
-// unblocks every queued request that no longer collides — they all start in
-// parallel rather than one-per-completion.
-func TestBaseRouter_QueueDrainPromotesMultiple(t *testing.T) {
-	a := newFakeProcess("a")
-	pb := newFakeProcess("b")
-	pc := newFakeProcess("c")
-
-	// A's swap evicts both B and C, so B and C must queue. Once A finishes
-	// B and C themselves have empty evict sets, so they can start together.
-	planner := &stubPlanner{evict: map[string][]string{
-		"a": {"b", "c"},
-	}}
-	b := newTestBase(t, map[string]process.Process{"a": a, "b": pb, "c": pc}, planner)
-
-	w1 := httptest.NewRecorder()
-	done1 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w1, newRequest("a"))
-		close(done1)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-	<-a.runStarted
-
-	// B and C arrive while A is loading. evict_b and evict_c are empty,
-	// but collidesWith returns true because they appear in A's evict set.
-	w2 := httptest.NewRecorder()
-	done2 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w2, newRequest("b"))
-		close(done2)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-
-	w3 := httptest.NewRecorder()
-	done3 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w3, newRequest("c"))
-		close(done3)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-
-	if got := pb.runCalls.Load(); got != 0 {
-		t.Errorf("b started early: runCalls=%d", got)
-	}
-	if got := pc.runCalls.Load(); got != 0 {
-		t.Errorf("c started early: runCalls=%d", got)
-	}
-
-	// Release A. The swapDone handler should drain the queue and start
-	// both B and C in parallel.
-	a.markReady()
-	waitProcessed(t, b.testProcessed, 1) // swapDone(A) → drainQueue starts B and C
-	<-pb.runStarted
-	<-pc.runStarted
-
-	pb.markReady()
-	pc.markReady()
-
-	for i, ch := range []chan struct{}{done1, done2, done3} {
-		select {
-		case <-ch:
-		case <-time.After(time.Second):
-			t.Fatalf("request %d did not complete", i)
-		}
-	}
-}
-
-// TestBaseRouter_Shutdown_FailsAllInFlight verifies that shutdown returns
-// the shutdown error to every waiter on every active swap AND to every
-// queued request.
-func TestBaseRouter_Shutdown_FailsAllInFlight(t *testing.T) {
-	a := newFakeProcess("a")
-	pb := newFakeProcess("b")
-	pc := newFakeProcess("c")
-
-	// a and b load in parallel (empty evicts). c collides with both.
-	planner := &stubPlanner{evict: map[string][]string{
-		"c": {"a", "b"},
-	}}
-	b := newTestBase(t, map[string]process.Process{"a": a, "b": pb, "c": pc}, planner)
-
-	const waitersPer = 2
-	var wg sync.WaitGroup
-	codes := make([]int, 0, 2*waitersPer+1)
-	var codesMu sync.Mutex
-	record := func(code int) {
-		codesMu.Lock()
-		codes = append(codes, code)
-		codesMu.Unlock()
-	}
-
-	launch := func(model string) {
-		wg.Add(1)
-		go func() {
-			defer wg.Done()
-			w := httptest.NewRecorder()
-			b.ServeHTTP(w, newRequest(model))
-			record(w.Code)
-		}()
-	}
-
-	// Active swaps for a and b, each with 2 waiters.
-	for i := 0; i < waitersPer; i++ {
-		launch("a")
-		waitProcessed(t, b.testProcessed, 1)
-	}
-	for i := 0; i < waitersPer; i++ {
-		launch("b")
-		waitProcessed(t, b.testProcessed, 1)
-	}
-	// c collides with both → queues.
-	launch("c")
-	waitProcessed(t, b.testProcessed, 1)
-
-	<-a.runStarted
-	<-pb.runStarted
-
-	if err := b.Shutdown(time.Second); err != nil {
-		t.Fatalf("Shutdown: %v", err)
-	}
-	wg.Wait()
-
-	codesMu.Lock()
-	defer codesMu.Unlock()
-	if len(codes) != 2*waitersPer+1 {
-		t.Fatalf("got %d responses, want %d", len(codes), 2*waitersPer+1)
-	}
-	for i, c := range codes {
-		if c == http.StatusOK {
-			t.Errorf("response %d: status=%d, want non-200 (shutdown)", i, c)
-		}
-	}
-}
-
-// TestBaseRouter_NoSwapWhileServing verifies that an already-loaded model
-// is not stopped to satisfy another model's swap while it is still handling
-// a request.
-//
-// Sequence:
-//  1. r1 (A) — A loads; ServeHTTP enters and is pinned via serveBlock.
-//  2. r2 (B, planner: B evicts A) — must NOT cause A.Stop while r1 is live.
-//  3. r3 (A) — arrives next; the existing code queues it because B's swap
-//     intent collides with A.
-//  4. r1 released — A finishes r1, then r3 is served by A.
-//  5. B's swap then proceeds; r2 is served by B.
-//
-// fakeProcess.stoppedWhileServing flips true if Stop is ever called while
-// a ServeHTTP is in flight — a direct, race-free signal of the violation.
-func TestBaseRouter_NoSwapWhileServing(t *testing.T) {
-	a := newFakeProcess("a")
-	// autoReady left false: we markReady manually after observing runStarted,
-	// so autoReady's setState(Ready) cannot race with a later Stop and leave
-	// A in Ready, masking the bug.
-	a.serveBlock = make(chan struct{})
-	pb := newFakeProcess("b")
-	// Same reasoning for B: park its swap on WaitReady until we choose.
-
-	planner := &stubPlanner{evict: map[string][]string{"b": {"a"}}}
-	b := newTestBase(t, map[string]process.Process{"a": a, "b": pb}, planner)
-
-	// r1 — load A and enter its ServeHTTP (which blocks on serveBlock).
-	w1 := httptest.NewRecorder()
-	done1 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w1, newRequest("a"))
-		close(done1)
-	}()
-	waitProcessed(t, b.testProcessed, 1) // handlerReq for r1
-	<-a.runStarted
-	a.markReady()
-	waitProcessed(t, b.testProcessed, 1) // swapDone for A
-	<-a.serveStarted
-
-	// r2 — would evict A. A must not be stopped while r1 is in flight.
-	w2 := httptest.NewRecorder()
-	done2 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w2, newRequest("b"))
-		close(done2)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-
-	// r3 — another request for A, arrives behind r2 and queues because
-	// B's swap intent (which evicts A) is recorded as active.
-	w3 := httptest.NewRecorder()
-	done3 := make(chan struct{})
-	go func() {
-		b.ServeHTTP(w3, newRequest("a"))
-		close(done3)
-	}()
-	waitProcessed(t, b.testProcessed, 1)
-
-	// Release r1 (and r3 if it is fast-pathed onto the still-loaded A).
-	// The router must hold off B's swap until A has drained.
-	close(a.serveBlock)
-
-	select {
-	case <-done1:
-	case <-time.After(2 * time.Second):
-		t.Fatal("r1 did not complete after serveBlock release")
-	}
-
-	// Wait for B.Run before marking it ready: markReady before Run would
-	// skip the Run path entirely and leave pb.runCalls at 0. In a correct
-	// implementation B's swap only starts after A has drained; in the
-	// current implementation it has already started — either way runStarted
-	// fires.
-	<-pb.runStarted
-	pb.markReady()
-
-	select {
-	case <-done2:
-	case <-time.After(2 * time.Second):
-		t.Fatal("r2 did not complete after B marked ready")
-	}
-	select {
-	case <-done3:
-	case <-time.After(2 * time.Second):
-		t.Fatal("r3 did not complete")
-	}
-
-	if w1.Code != http.StatusOK || w2.Code != http.StatusOK || w3.Code != http.StatusOK {
-		t.Fatalf("statuses: w1=%d w2=%d w3=%d", w1.Code, w2.Code, w3.Code)
-	}
-	if w1.Body.String() != "ok:a" {
-		t.Errorf("r1 body=%q want ok:a", w1.Body.String())
-	}
-	if w3.Body.String() != "ok:a" {
-		t.Errorf("r3 body=%q want ok:a (r3 must be served by A)", w3.Body.String())
-	}
-	if w2.Body.String() != "ok:b" {
-		t.Errorf("r2 body=%q want ok:b", w2.Body.String())
-	}
-	if a.stoppedWhileServing.Load() {
-		t.Errorf("A.Stop was called while A was still handling a request — the router swapped out a busy process")
-	}
-}
-
 func TestBaseRouter_ModelNotFound(t *testing.T) {
 	a := newFakeProcess("a")
 	b := newTestBase(t, map[string]process.Process{"a": a}, &stubPlanner{})

internal/router/design.md

@@ -0,0 +1,404 @@
+# Router design
+
+A developer tutorial for the `internal/router` package and its `scheduler`
+sub-package.
+
+## Intro
+
+A llama-swap router is the component that sits behind the proxy and answers one
+question for every incoming request: _can this model serve right now, and if
+not, what has to happen first?_ Answering it means juggling three concerns that
+used to live tangled together in one type:
+
+1. **Process machinery** — owning the OS processes, starting and stopping them,
+   running health checks, and shuttling HTTP requests onto the right upstream.
+2. **Scheduling strategy** — the queue, in-flight bookkeeping, and the decision
+   tree that turns one request into "serve now", "join an existing swap",
+   "queue", or "start a swap".
+3. **Eviction policy** — given a model we want to load, which currently-running
+   models have to be stopped to make room?
+
+The design pulls those three apart into separate, independently replaceable
+pieces:
+
+| Concern             | Type                           | Lives in                        |
+| ------------------- | ------------------------------ | ------------------------------- |
+| Process machinery   | `baseRouter`                   | `internal/router/base.go`       |
+| Scheduling strategy | `scheduler.Scheduler` (`FIFO`) | `internal/router/scheduler/`    |
+| Eviction policy     | `scheduler.Swapper`            | `groupSwapper`, `matrixSwapper` |
+
+`baseRouter` keeps the channels, run loop, process lifecycle, and shutdown
+teardown, and exposes the side-effects a scheduler needs through the
+`scheduler.Effects` interface. The scheduler owns the queue and decision tree
+but performs no side-effects directly — it calls back through `Effects`. The
+`Swapper` is a pure function from "target model + currently running" to "models
+to evict", and knows nothing about queues, channels, or processes.
+
+Because the seams are interfaces, you can replace the scheduling strategy
+without touching process management, or write a new eviction policy without
+touching either. `FIFO` is the first and currently only `Scheduler`;
+`groupSwapper` and `matrixSwapper` are the two `Swapper`s.
+
+## Key concepts
+
+### One run loop, no locks
+
+`baseRouter.run()` is a single goroutine selecting over a handful of channels:
+
+```go
+for {
+    select {
+    case req := <-b.shutdownCh:  b.handleShutdown(req); return
+    case req := <-b.handlerCh:   b.schedule.OnRequest(req)
+    case req := <-b.unloadCh:    b.schedule.OnUnload(req.targets, req.timeout); close(req.respond)
+    case ev := <-b.swapDoneCh:   b.schedule.OnSwapDone(ev)
+    case ev := <-b.serveDoneCh:  b.schedule.OnServeDone(ev)
+    }
+}
+```
+
+Every `Scheduler` method runs on this one goroutine. That is the single most
+important fact about the design: **the scheduler never needs a mutex for its own
+state**. All scheduler state is touched only from these callbacks, which are
+serialized by the run loop. If you write a new scheduler, you get the same
+guarantee for free — and you must not break it by spinning up goroutines that
+mutate scheduler state.
+
+### Events flow in, side-effects flow out
+
+The run loop turns external happenings into method calls on the scheduler:
+
+- A new HTTP request becomes `OnRequest(HandlerReq)`.
+- A swap goroutine finishing becomes `OnSwapDone(SwapDone)`.
+- A tracked request handler returning becomes `OnServeDone(ServeDoneEvent)`.
+- An admin unload becomes `OnUnload(targets, timeout)`.
+- Shutdown becomes `OnShutdown(err)`.
+
+The scheduler reacts by calling **back out** through `Effects`: inspect a
+process state, start a swap, grant a response to a caller, or stop processes. It
+never calls `process.Process` directly and never writes to a channel directly.
+This keeps the scheduler pure enough to unit-test against a fake `Effects` with
+no goroutines or real processes involved (see `scheduler/fifo_test.go`).
+
+```
+   HTTP request                          admin Unload / Shutdown
+        │                                          │
+        ▼                                          ▼
+ ServeHTTP ──HandlerReq──▶  baseRouter.run()  ◀──unloadCh/shutdownCh
+                                  │  (single goroutine)
+                                  ▼
+                          Scheduler.On*(...)
+                                  │  calls back through
+                                  ▼
+                          Effects: ModelState / StartSwap /
+                                   GrantServe / GrantError / StopProcesses
+                                  │
+                                  ▼
+                  baseRouter side-effects: doSwap goroutine,
+                  grant() to caller, process.Stop()
+                                  │
+            swap completes ──SwapDone──▶ back into run loop
+```
+
+### The swap goroutine
+
+Scheduling decisions must be quick and non-blocking, but loading a model is
+slow. The two are reconciled by doing the slow part on a separate goroutine.
+
+When the scheduler decides to start a swap, inside `OnRequest` it:
+
+1. records "a swap for X is in flight" in its own state, then
+2. calls `Effects.StartSwap(modelID, evict)`.
+
+`StartSwap` does **not** load the model itself — it just launches a detached
+goroutine (`doSwap`) and returns straight away. `doSwap` is what does the slow
+work: stop the evicted processes, start the target, wait for it to become ready.
+Because `StartSwap` returned immediately, `OnRequest` returns too, and the run
+loop is free to pick up the next event — another request, a serve-done, an
+unload — while `doSwap` runs in the background.
+
+The swap's eventual result comes back as just another event: when `doSwap`
+finishes it posts a `SwapDone` onto `swapDoneCh`, which the run loop delivers as
+`OnSwapDone`. So a slow load never blocks the run loop; it brackets it with two
+quick events (`OnRequest` to start, `OnSwapDone` to finish) and everything in
+between is handled normally.
+
+### In-flight tracking and `trackedServe`
+
+When the scheduler grants a request, the handler it hands back is wrapped by
+`baseRouter.trackedServe`. The wrapper runs the real `ServeHTTP` and, on return,
+posts a `ServeDoneEvent` so the run loop can decrement the per-model in-flight
+count. This is why the scheduler can know whether a process is "busy": it counts
+grants out and serve-dones in. A swap that would evict a busy process is
+deferred until that process's in-flight count hits zero (`OnServeDone` then
+re-drains the queue).
+
+The subtle contract here is `GrantServe`'s boolean return. The caller's
+`Respond` channel is unbuffered, so a successful send proves the HTTP goroutine
+is alive and took the handler. If the caller already disconnected, the send
+fails, `trackedServe` never runs, and **no** `ServeDoneEvent` will ever arrive —
+so the scheduler must only increment `inFlight` when `GrantServe` returns true.
+Incrementing on a false return would strand the counter above zero and the model
+could never be evicted again.
+
+## The interfaces
+
+All three live in `scheduler/scheduler.go`.
+
+### `Scheduler`
+
+```go
+type Scheduler interface {
+    OnRequest(req HandlerReq)
+    OnSwapDone(ev SwapDone)
+    OnServeDone(ev ServeDoneEvent)
+    OnUnload(targets []string, timeout time.Duration)
+    OnShutdown(err error)
+}
+```
+
+Owns the queue, in-flight tracking, and the decision tree. All methods run on
+the run-loop goroutine, so no internal locking is needed.
+
+### `Swapper`
+
+```go
+type Swapper interface {
+    EvictionFor(target string, running []string) []string
+    OnSwapStart(target string, running []string)
+}
+```
+
+The eviction policy. `EvictionFor` is a **pure decision** — given the target and
+the complete `running` set, return the running model IDs that must stop. It must
+not log or mutate anything, and it does **not** inspect process state itself:
+the scheduler hands it `running` already assembled (every non-stopped process,
+unioned with the targets of in-flight swaps already committed but not yet
+visible in process state). That keeps the swapper a pure function of its inputs,
+with no reference to processes.
+
+The reason it must not log is that it is a _speculative_ query — "what would we
+evict if we started this swap right now?" — called far more often than swaps
+actually happen. The scheduler calls it once per incoming request, and then
+**again for every still-queued request on every queue drain** (each `OnSwapDone`,
+`OnServeDone`, and `OnUnload`). Most of those calls end in "still queued",
+"collides", or "nothing to evict", not a real swap. Logging there would emit
+duplicate lines for a request that simply sits in the queue, and lines for
+decisions that never happen — the log would stop meaning "a swap occurred".
+
+`OnSwapStart` is the one place a Swapper may log, because it is called exactly
+once, at the moment a swap is committed. One log line there equals one real swap,
+with the evict set that is genuinely being applied — which is why `matrixSwapper`
+re-solves and logs the full decision (set, DSL, cost) in `OnSwapStart` rather
+than in `EvictionFor`.
+
+### `Effects`
+
+```go
+type Effects interface {
+    ModelState(modelID string) (process.ProcessState, bool)
+    RunningModels() map[string]process.ProcessState
+    StartSwap(modelID string, evict []string)
+    GrantError(req HandlerReq, err error)
+    GrantServe(req HandlerReq, modelID string) bool
+    StopProcesses(timeout time.Duration, ids []string)
+}
+```
+
+Implemented by `baseRouter`. This is the scheduler's entire window onto the
+outside world; everything else about the router is hidden from it. See the
+deep-dive below.
+
+### `Factory` — wiring it together
+
+```go
+type Factory func(name string, logger *logmon.Monitor, eff Effects) Scheduler
+```
+
+`baseRouter` doesn't know which scheduler or swapper it has — it is handed a
+`Factory` at construction and calls it once, passing itself as the `Effects`.
+The concrete router captures its `Swapper` in the closure. From `group.go`:
+
+```go
+swapper := &groupSwapper{ /* ... */ }
+base := newBaseRouter("group", conf, processes, proxylog,
+    func(name string, logger *logmon.Monitor, eff scheduler.Effects) scheduler.Scheduler {
+        return scheduler.NewFIFO(name, logger, swapper, eff)
+    })
+```
+
+This closure is the single point where the three pieces meet: it binds a
+specific `Swapper` (`swapper`) and a specific `Scheduler` (`FIFO`) to the
+`baseRouter`'s `Effects` (`eff`).
+
+**The swapper is a separate type from the concrete router.** There are currently two router implementations router.Group and router.Matrix. Each of these has a custom swapper that implements scheduler.Swapper for custom eviction logic. This decoupling of responsibilities makes it easy to implement custom swapping strategies.
+
+### The events
+
+A single goroutine in `baseRouter.run()` owns and serializes all state changes in the router. By processing events one at a time it ensures correctness and eliminates complex mutex lock logic.
+
+These are the events the router currently uses:
+
+```go
+type HandlerReq struct {            // one in-flight ServeHTTP awaiting a decision
+    Model      string
+    Ctx        context.Context
+    Respond    chan HandlerResp     // UNBUFFERED — see GrantServe contract
+    PositionCh chan int             // queue-position updates for the loading UI
+}
+
+type HandlerResp struct {           // the decision handed back to the caller
+    HandleFunc http.HandlerFunc     // serve with this, or...
+    Err        error                // ...fail with this
+}
+
+type SwapDone        struct{ ModelID string; Err error } // swap goroutine finished
+type ServeDoneEvent  struct{ ModelID string }            // tracked handler returned
+```
+
+## Deep-dive: the `Effects` interface and why it exists
+
+`Effects` is the inversion-of-control boundary that makes the split possible.
+The scheduler decides and `baseRouter` _acts_. Pulling the side-effects behind this
+interface buys three things:
+
+1. **Purity and testability.** The scheduler performs no I/O, starts no
+   goroutines of its own, and touches no real processes. Its tests drive the
+   `On*` methods directly and assert on a `fakeEffects` that just records the
+   calls — synchronous, deterministic, no sleeps. (`scheduler/fifo_test.go`.)
+2. **A single, auditable side-effect surface.** Every externally-visible thing a
+   scheduler can do is one of six methods. You can reason about the whole
+   contract by reading one interface.
+3. **Decoupling lifetime.** The scheduler never holds a `process.Process`,
+   never sees a channel, and never learns how shutdown teardown works. It only
+   knows model IDs and states.
+
+Method by method, as implemented in `base.go`:
+
+- **`ModelState(modelID) (state, ok)`** — read-only snapshot of a process's
+  state, and whether this router handles the model at all. The scheduler uses it
+  for the "unknown model" check and the "already ready" fast path. Safe to call
+  any time because the process map is fixed at construction and `State()` is a
+  snapshot.
+
+- **`RunningModels()`** — the state of every process that isn't stopped or shut
+  down. The scheduler unions its keys with its own in-flight swap targets to
+  build the `running` set it hands the `Swapper`, so the swapper never has to
+  touch process state itself.
+
+- **`StartSwap(modelID, evict)`** — fire-and-forget. `baseRouter` launches the
+  `doSwap` goroutine and returns immediately; the result comes back later as a
+  `SwapDone`. The scheduler records the swap as active _before_ calling this so
+  that requests arriving in the meantime can join it.
+
+- **`GrantError(req, err)`** — hand a caller an error response. Used for unknown
+  models, failed swaps, unloads, and shutdown.
+
+- **`GrantServe(req, modelID) bool`** — hand a caller the tracked handler for a
+  ready model, returning whether the caller was still there to receive it. The
+  scheduler increments the in-flight count **only on a true return** (see the
+  in-flight contract above). This is the one `Effects` method whose return value
+  carries state-machine significance.
+
+- **`StopProcesses(timeout, ids)`** — stop processes in parallel and **block**
+  until all have stopped. Used by `OnUnload` so an admin `Unload` call can
+  guarantee the process is dead by the time it returns. (Note `StartSwap` is
+  async but `StopProcesses` is sync — the difference is deliberate and tied to
+  the caller's expectations.)
+
+A useful way to hold it in your head: `Effects` is the scheduler's syscall
+table. The scheduler is a pure state machine; `Effects` is how it touches the
+world, and `baseRouter` is the kernel that implements those syscalls with real
+goroutines, channels, and processes.
+
+## How to implement a new `Swapper`
+
+A `Swapper` is a pure decision function plus a logging hook — the easiest of the three pieces to replace.
+
+1. **Write the swapper type** and give it whatever config it needs to make a
+   decision. It does **not** need the process map — the scheduler supplies the
+   running set as an argument. `groupSwapper` holds only its group config;
+   `matrixSwapper` holds only its solver and logger:
+
+   ```go
+   type mySwapper struct {
+       config config.Config
+   }
+   ```
+
+2. **Implement `EvictionFor(target, running)`** as a _pure_ decision:
+   - `running` is the complete live set, already assembled for you: every
+     non-stopped process unioned with the targets of in-flight swaps the
+     scheduler has committed to. You don't filter process state or fold in
+     in-flight targets yourself, that's the scheduler's job. Just decide against the slice you're handed.
+   - Return the list of model IDs in `running` that must stop for `target` to
+     run. Return `nil`/empty when nothing needs evicting.
+   - Do **not** mutate state here.
+   - Do **not** log here. It can be called multiple times per request. Since it is pure function have tests verify the expected behaviour.
+
+3. **Implement `OnSwapStart(target, running)`** — called once when a swap
+   actually begins, with the same `running` set `EvictionFor` saw. This is the
+   right place to log: one call equals one real swap. `matrixSwapper` re-solves
+   and logs the chosen set and cost here; `groupSwapper` logs nothing.
+
+4. **Wire it in** by instantiating the swapper in your router's constructor and
+   capturing it in the `Factory` closure passed to `newBaseRouter` — exactly as
+   `NewGroup` and `NewMatrix` do. The router struct itself only ever embeds
+   `*baseRouter`; the swapper reaches the scheduler solely through that closure.
+
+Reference implementations: `groupSwapper` (static group config) in `group.go`
+and `matrixSwapper` (cost-based set solver) in `matrix.go`.
+
+## How to implement a new `Scheduler`
+
+Replacing the scheduler means taking over the queue and the entire decision tree. Read `scheduler/fifo.go` end to end first — it is the reference implementation and the rules below are easiest to understand in context.
+
+The rules you must honour:
+
+- **Single goroutine.** Every method runs on the `baseRouter.run()` goroutine. Keep your state in plain maps/slices and never read or write it from another goroutine. If you need slow work done, hand it to `Effects.StartSwap` and react to the resulting `SwapDone` — do not block a method waiting for it.
+
+- **Never block the run loop.** `OnRequest`, `OnSwapDone`, and `OnServeDone` must make a decision and return. The one method allowed to block is `OnUnload`, and only because it must wait on the synchronous `StopProcesses` so the admin caller's guarantee holds.
+
+- **Respect the `GrantServe` boolean.** Only count a request as in-flight when `GrantServe` returns true (see the in-flight contract above). A false return means the caller is gone; no `ServeDoneEvent` will ever arrive, so incrementing on false permanently strands the counter.
+
+- **Account for in-flight swaps in your running set.** When you call `Swapper.EvictionFor`, the running set you pass must include not just live processes (`Effects.RunningModels`) but also the targets of swaps you've already started that aren't yet visible in process state — otherwise the swapper contradicts decisions already in motion.
+
+What each method must do:
+
+- **`OnRequest(req)`** — every request must resolve to exactly one of: granted, errored, joined (piggybacks an in-flight swap), queued, or swap-started. No request may be silently dropped.
+
+- **`OnSwapDone(ev)`** — deliver the result to every waiter that joined this swap (grant on success, error on `ev.Err`), drop the swap from active tracking, then re-examine anything queued — a finished swap may have unblocked it.
+
+- **`OnServeDone(ev)`** — decrement the model's in-flight count; when it hits zero, re-examine the queue. Do **not** clear in-flight counts by hand; the handlers post their own `ServeDoneEvent`s on return.
+
+- **`OnUnload(targets, timeout)`** — error out any waiters or queued requests for the unloaded models, call `Effects.StopProcesses` (synchronously — the admin caller relies on the process being dead afterwards), then re-examine the queue.
+
+- **`OnShutdown(err)`** — error out every waiter you still hold (active swap waiters and queued requests). Don't touch processes; teardown is `baseRouter`'s job.
+
+Expose a constructor matching the `Factory` shape:
+
+```go
+func NewMyScheduler(name string, logger *logmon.Monitor, swapper Swapper, eff Effects) *MyScheduler {
+    // ...
+}
+
+// in the concrete router:
+base := newBaseRouter(name, conf, processes, proxylog,
+    func(name string, logger *logmon.Monitor, eff scheduler.Effects) scheduler.Scheduler {
+        return scheduler.NewMyScheduler(name, logger, swapper, eff)
+    })
+```
+
+## Testing
+
+- **Schedulers** are tested as pure state machines in the `scheduler` package:
+  drive the `On*` methods directly against a `fakeEffects` and assert on the
+  recorded grants/starts/stops. No goroutines, no sleeps. See
+  `scheduler/fifo_test.go` as the reference; follow the `TestSchedulerName_<scenario>`
+  naming convention.
+- **`baseRouter` mechanism** (run loop, `grant`/`ServeHTTP`, `Unload`,
+  `Shutdown`) is tested in `base_test.go`. The run loop exposes a
+  `testProcessed` channel so tests can wait for an event to be fully processed
+  instead of sleeping.
+- Run new tests with `go test -v -run TestMyScheduler_... ./internal/router/scheduler/`,
+  then `make test-dev` for a quick `go test` + `staticcheck` pass over `proxy/`.

internal/router/group.go

@@ -14,7 +14,7 @@ type Group struct {
 
 func NewGroup(conf config.Config, proxylog, upstreamlog *logmon.Monitor) (*Group, error) {
 	modelToGroup := make(map[string]string)
-	for gid, gcfg := range conf.Groups {
+	for gid, gcfg := range conf.Routing.Router.Settings.Groups {
 		for _, mid := range gcfg.Members {
 			if existing, dup := modelToGroup[mid]; dup {
 				return nil, fmt.Errorf("model %q is in multiple groups: %q and %q", mid, existing, gid)
@@ -23,25 +23,29 @@ func NewGroup(conf config.Config, proxylog, upstreamlog *logmon.Monitor) (*Group
 		}
 	}
 
-	planner := &groupPlanner{
+	swapper := &groupSwapper{
 		config:       conf,
 		modelToGroup: modelToGroup,
 	}
 
 	processes := make(map[string]process.Process, len(modelToGroup))
-	base := newBaseRouter("group", conf, processes, planner, proxylog)
-	planner.processes = processes
+	base, err := newBaseRouter("group", conf, processes, proxylog, swapper)
+	if err != nil {
+		return nil, fmt.Errorf("creating base router: %w", err)
+	}
 
 	for mid := range modelToGroup {
 		modelCfg, _, ok := conf.FindConfig(mid)
 		if !ok {
 			base.shutdownFn()
+			base.procCancel()
 			return nil, fmt.Errorf("no model config for %q", mid)
 		}
 		procLog := logmon.NewWriter(upstreamlog)
-		p, err := process.New(base.shutdownCtx, mid, modelCfg, procLog, proxylog)
+		p, err := process.New(base.procCtx, mid, modelCfg, procLog, proxylog)
 		if err != nil {
 			base.shutdownFn()
+			base.procCancel()
 			return nil, fmt.Errorf("creating process for %q: %w", mid, err)
 		}
 		processes[mid] = p
@@ -52,21 +56,20 @@ func NewGroup(conf config.Config, proxylog, upstreamlog *logmon.Monitor) (*Group
 	return g, nil
 }
 
-// groupPlanner decides evictions from static group configuration.
+// groupSwapper decides evictions from static group configuration.
 //
 // Same-group siblings are stopped when the group has swap=true. Cross-group
 // members are stopped only when the target's group is exclusive; loading a
 // model from a non-exclusive group leaves running exclusive groups alone,
 // matching the gotcha in the original ProcessGroup behaviour.
-type groupPlanner struct {
+type groupSwapper struct {
 	config       config.Config
 	modelToGroup map[string]string
-	processes    map[string]process.Process
 }
 
-func (p *groupPlanner) EvictionFor(target string, alsoRunning []string) []string {
+func (p *groupSwapper) EvictionFor(target string, running []string) []string {
 	tg := p.modelToGroup[target]
-	tgCfg := p.config.Groups[tg]
+	tgCfg := p.config.Routing.Router.Settings.Groups[tg]
 
 	seen := make(map[string]struct{})
 	var result []string
@@ -87,24 +90,17 @@ func (p *groupPlanner) EvictionFor(target string, alsoRunning []string) []string
 		// for backwards compatibility. The newer swap matrix approach does not
 		// have this issue.
 		case og != tg && tgCfg.Exclusive:
-			if ogCfg := p.config.Groups[og]; !ogCfg.Persistent {
+			if ogCfg := p.config.Routing.Router.Settings.Groups[og]; !ogCfg.Persistent {
 				seen[mID] = struct{}{}
 				result = append(result, mID)
 			}
 		}
 	}
 
-	for mID, proc := range p.processes {
-		st := proc.State()
-		if st == process.StateStopped || st == process.StateShutdown {
-			continue
-		}
-		consider(mID)
-	}
-	for _, mID := range alsoRunning {
+	for _, mID := range running {
 		consider(mID)
 	}
 	return result
 }
 
-func (p *groupPlanner) OnSwapStart(target string) {}
+func (p *groupSwapper) OnSwapStart(target string, running []string) {}

internal/router/group_test.go

@@ -17,17 +17,19 @@ import (
 func newTestGroup(t *testing.T, conf config.Config, processes map[string]process.Process) *Group {
 	t.Helper()
 	modelToGroup := make(map[string]string)
-	for gid, gcfg := range conf.Groups {
+	for gid, gcfg := range conf.Routing.Router.Settings.Groups {
 		for _, mid := range gcfg.Members {
 			modelToGroup[mid] = gid
 		}
 	}
-	planner := &groupPlanner{
+	swapper := &groupSwapper{
 		config:       conf,
 		modelToGroup: modelToGroup,
-		processes:    processes,
 	}
-	base := newBaseRouter("group", conf, processes, planner, logmon.NewWriter(io.Discard))
+	base, err := newBaseRouter("group", conf, processes, logmon.NewWriter(io.Discard), swapper)
+	if err != nil {
+		t.Fatalf("newBaseRouter: %v", err)
+	}
 	base.testProcessed = make(chan struct{}, 64)
 	g := &Group{baseRouter: base}
 	go base.run()
@@ -41,10 +43,10 @@ func newTestGroup(t *testing.T, conf config.Config, processes map[string]process
 
 func TestGroup_NewGroup_DuplicateMembership(t *testing.T) {
 	conf := config.Config{
-		Groups: map[string]config.GroupConfig{
+		Routing: groupRouting(map[string]config.GroupConfig{
 			"g1": {Swap: true, Members: []string{"a"}},
 			"g2": {Swap: true, Members: []string{"a"}},
-		},
+		}),
 		Models: map[string]config.ModelConfig{
 			"a": {},
 		},
@@ -65,9 +67,9 @@ func TestGroup_ServeHTTP_SwapStopsPrevious(t *testing.T) {
 
 	conf := config.Config{
 		HealthCheckTimeout: 5,
-		Groups: map[string]config.GroupConfig{
+		Routing: groupRouting(map[string]config.GroupConfig{
 			"g": {Swap: true, Exclusive: true, Members: []string{"a", "b"}},
-		},
+		}),
 	}
 	g := newTestGroup(t, conf, map[string]process.Process{"a": a, "b": b})
 
@@ -97,9 +99,9 @@ func TestGroup_NonSwapGroup_NoStop(t *testing.T) {
 
 	conf := config.Config{
 		HealthCheckTimeout: 5,
-		Groups: map[string]config.GroupConfig{
+		Routing: groupRouting(map[string]config.GroupConfig{
 			"g": {Swap: false, Exclusive: false, Members: []string{"a", "b"}},
-		},
+		}),
 	}
 	g := newTestGroup(t, conf, map[string]process.Process{"a": a, "b": b})
 
@@ -127,10 +129,10 @@ func TestGroup_CrossGroupExclusive(t *testing.T) {
 
 	conf := config.Config{
 		HealthCheckTimeout: 5,
-		Groups: map[string]config.GroupConfig{
+		Routing: groupRouting(map[string]config.GroupConfig{
 			"g1": {Swap: true, Exclusive: true, Members: []string{"a"}},
 			"g2": {Swap: true, Exclusive: true, Members: []string{"b"}},
-		},
+		}),
 	}
 	g := newTestGroup(t, conf, map[string]process.Process{"a": a, "b": b})
 
@@ -154,10 +156,10 @@ func TestGroup_CrossGroupNonExclusiveParallel(t *testing.T) {
 
 	conf := config.Config{
 		HealthCheckTimeout: 5,
-		Groups: map[string]config.GroupConfig{
+		Routing: groupRouting(map[string]config.GroupConfig{
 			"g1": {Swap: true, Exclusive: false, Members: []string{"a"}},
 			"g2": {Swap: true, Exclusive: false, Members: []string{"b"}},
-		},
+		}),
 	}
 	g := newTestGroup(t, conf, map[string]process.Process{"a": a, "b": pb})
 
@@ -202,16 +204,17 @@ func TestGroup_CrossGroupNonExclusiveParallel(t *testing.T) {
 
 // TestGroup_SameGroupSwapSerialises verifies that two same-group requests
 // (Swap=true) serialise even when both arrive while neither has reached
-// StateStarting yet — the alsoRunning hint to the planner closes that race.
+// StateStarting yet — the in-flight swap target the scheduler folds into the
+// running set closes that race.
 func TestGroup_SameGroupSwapSerialises(t *testing.T) {
 	a := newFakeProcess("a")
 	pb := newFakeProcess("b")
 
 	conf := config.Config{
 		HealthCheckTimeout: 5,
-		Groups: map[string]config.GroupConfig{
+		Routing: groupRouting(map[string]config.GroupConfig{
 			"g": {Swap: true, Exclusive: false, Members: []string{"a", "b"}},
-		},
+		}),
 	}
 	g := newTestGroup(t, conf, map[string]process.Process{"a": a, "b": pb})
 
@@ -224,8 +227,9 @@ func TestGroup_SameGroupSwapSerialises(t *testing.T) {
 	waitProcessed(t, g.testProcessed, 1)
 
 	// Request B arrives before A transitions to StateStarting in the process
-	// state machine. Without the alsoRunning hint, the planner would not see
-	// A as running, and B would start in parallel, violating Swap=true.
+	// state machine. Without folding the in-flight swap target into the running
+	// set, the swapper would not see A as running, and B would start in
+	// parallel, violating Swap=true.
 	w2 := httptest.NewRecorder()
 	done2 := make(chan struct{})
 	go func() {
@@ -269,10 +273,10 @@ func TestGroup_PersistentNotEvicted(t *testing.T) {
 
 	conf := config.Config{
 		HealthCheckTimeout: 5,
-		Groups: map[string]config.GroupConfig{
+		Routing: groupRouting(map[string]config.GroupConfig{
 			"persist": {Swap: true, Exclusive: false, Persistent: true, Members: []string{"a"}},
 			"other":   {Swap: true, Exclusive: true, Members: []string{"b"}},
-		},
+		}),
 	}
 	g := newTestGroup(t, conf, map[string]process.Process{"a": a, "b": b})
 
@@ -306,10 +310,10 @@ func TestGroup_NonExclusiveDoesNotUnloadExclusive(t *testing.T) {
 
 	conf := config.Config{
 		HealthCheckTimeout: 5,
-		Groups: map[string]config.GroupConfig{
+		Routing: groupRouting(map[string]config.GroupConfig{
 			"g1": {Swap: true, Exclusive: true, Members: []string{"a"}},
 			"g2": {Swap: true, Exclusive: false, Members: []string{"b"}},
-		},
+		}),
 	}
 	g := newTestGroup(t, conf, map[string]process.Process{"a": a, "b": b})
 

internal/router/helpers_test.go

@@ -12,10 +12,23 @@ import (
 	"testing"
 	"time"
 
+	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
 	"github.com/mostlygeek/llama-swap/internal/process"
 )
 
+// groupRouting builds a normalized RoutingConfig for the group router, mirroring
+// what config.LoadConfigFromReader produces. Tests use it to populate
+// config.Config.Routing without going through LoadConfig.
+func groupRouting(groups map[string]config.GroupConfig) config.RoutingConfig {
+	return config.RoutingConfig{
+		Router: config.RouterConfig{
+			Use:      "group",
+			Settings: config.RouterSettings{Groups: groups},
+		},
+	}
+}
+
 // fakeProcess is an in-memory implementation of process.Process used to drive
 // the routers through their state machine without spawning real upstreams.
 type fakeProcess struct {

internal/router/loading.go

@@ -38,6 +38,13 @@ type loadingWriter struct {
 	pendingMu     sync.Mutex
 	pendingUpdate string
 
+	// writeMu serializes writes to the underlying writer and guards released.
+	// Once released is set, the streaming goroutine must not touch the writer
+	// again — ServeHTTP has reclaimed it (to run the real handler or to return)
+	// and writing/flushing a finalized response panics.
+	writeMu  sync.Mutex
+	released bool
+
 	// closed by start when the goroutine finishes (after cleanup messages)
 	done chan struct{}
 
@@ -217,12 +224,33 @@ func (s *loadingWriter) sendData(data string) {
 		return
 	}
 
-	_, err = fmt.Fprintf(s.writer, "data: %s\n\n", jsonData)
-	if err != nil {
+	s.writeMu.Lock()
+	defer s.writeMu.Unlock()
+	// Once ServeHTTP has reclaimed the writer (release), writing/flushing it
+	// races the real handler or panics on a finalized response. Stop here.
+	if s.released {
+		return
+	}
+
+	if _, err = fmt.Fprintf(s.writer, "data: %s\n\n", jsonData); err != nil {
 		s.logger.Debugf("<%s> Failed to write SSE data (client likely disconnected): %v", s.modelName, err)
 		return
 	}
-	s.Flush()
+	if flusher, ok := s.writer.(http.Flusher); ok {
+		flusher.Flush()
+	}
+}
+
+// release fences the loadingWriter off from the underlying ResponseWriter.
+// After it returns, the streaming goroutine will not write to or flush the
+// writer again: any in-flight write completes under writeMu first, and later
+// writes short-circuit on released. The caller can then safely hand the writer
+// to the real handler or let ServeHTTP return without racing a finalized
+// response (a use-after-return Flush panics on the recycled *bufio.Writer).
+func (s *loadingWriter) release() {
+	s.writeMu.Lock()
+	s.released = true
+	s.writeMu.Unlock()
 }
 
 func (s *loadingWriter) Header() http.Header {

internal/router/loading_test.go

@@ -226,69 +226,6 @@ func TestIsLoadingPath(t *testing.T) {
 	}
 }
 
-func TestExtractContext_Streaming_GET(t *testing.T) {
-	tests := []struct {
-		name          string
-		query         string
-		wantStreaming bool
-	}{
-		{"streaming true", "model=llama3&stream=true", true},
-		{"streaming false", "model=llama3&stream=false", false},
-		{"no stream param", "model=llama3", false},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			r, _ := http.NewRequest(http.MethodGet, "/?"+tt.query, nil)
-			got, err := ExtractContext(r)
-			if err != nil {
-				t.Fatalf("ExtractContext: %v", err)
-			}
-			if got.Streaming != tt.wantStreaming {
-				t.Errorf("Streaming: want %v, got %v", tt.wantStreaming, got.Streaming)
-			}
-		})
-	}
-}
-
-func TestExtractContext_Streaming_JSON(t *testing.T) {
-	tests := []struct {
-		name          string
-		body          string
-		wantStreaming bool
-	}{
-		{"streaming true", `{"model":"llama3","stream":true}`, true},
-		{"streaming false", `{"model":"llama3","stream":false}`, false},
-		{"no stream param", `{"model":"llama3"}`, false},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			r, _ := http.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(tt.body))
-			r.Header.Set("Content-Type", "application/json")
-			got, err := ExtractContext(r)
-			if err != nil {
-				t.Fatalf("ExtractContext: %v", err)
-			}
-			if got.Streaming != tt.wantStreaming {
-				t.Errorf("Streaming: want %v, got %v", tt.wantStreaming, got.Streaming)
-			}
-		})
-	}
-}
-
-func TestExtractContext_Streaming_URLEncodedForm(t *testing.T) {
-	r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", strings.NewReader("model=whisper-1&stream=true"))
-	r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
-	got, err := ExtractContext(r)
-	if err != nil {
-		t.Fatalf("ExtractContext: %v", err)
-	}
-	if !got.Streaming {
-		t.Error("Streaming should be true")
-	}
-}
-
 func countSSEMessages(s string) int {
 	scanner := bufio.NewScanner(strings.NewReader(s))
 	count := 0

internal/router/matrix.go

@@ -2,7 +2,6 @@ package router
 
 import (
 	"fmt"
-	"sort"
 
 	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
@@ -14,26 +13,30 @@ type Matrix struct {
 }
 
 func NewMatrix(conf config.Config, proxylog, upstreamlog *logmon.Monitor) (*Matrix, error) {
-	if conf.Matrix == nil {
+	mtx := conf.Routing.Router.Settings.Matrix
+	if mtx == nil {
 		return nil, fmt.Errorf("matrix router requires a matrix configuration")
 	}
 
-	planner := &matrixPlanner{
-		solver: newMatrixSolver(conf.ExpandedSets, conf.Matrix.ResolvedEvictCosts()),
+	swapper := &matrixSwapper{
+		solver: newMatrixSolver(mtx.ExpandedSets, mtx.ResolvedEvictCosts()),
 		logger: proxylog,
 	}
 
 	// Build a process for every model in the config. Any model can run alone
 	// even if it is not part of a set; this mirrors proxy.NewMatrix.
 	processes := make(map[string]process.Process, len(conf.Models))
-	base := newBaseRouter("matrix", conf, processes, planner, proxylog)
-	planner.processes = processes
+	base, err := newBaseRouter("matrix", conf, processes, proxylog, swapper)
+	if err != nil {
+		return nil, fmt.Errorf("creating base router: %w", err)
+	}
 
 	for mid, modelCfg := range conf.Models {
 		procLog := logmon.NewWriter(upstreamlog)
-		p, err := process.New(base.shutdownCtx, mid, modelCfg, procLog, proxylog)
+		p, err := process.New(base.procCtx, mid, modelCfg, procLog, proxylog)
 		if err != nil {
 			base.shutdownFn()
+			base.procCancel()
 			return nil, fmt.Errorf("creating process for %q: %w", mid, err)
 		}
 		processes[mid] = p
@@ -44,20 +47,18 @@ func NewMatrix(conf config.Config, proxylog, upstreamlog *logmon.Monitor) (*Matr
 	return r, nil
 }
 
-// matrixPlanner decides evictions by asking the matrix solver against the
-// current running set.
-type matrixPlanner struct {
-	solver    *matrixSolver
-	processes map[string]process.Process
-	logger    *logmon.Monitor
+// matrixSwapper decides evictions by asking the matrix solver against the
+// running set the scheduler hands it.
+type matrixSwapper struct {
+	solver *matrixSolver
+	logger *logmon.Monitor
 }
 
-func (p *matrixPlanner) EvictionFor(target string, alsoRunning []string) []string {
-	return p.solver.Solve(target, p.runningSet(alsoRunning)).Evict
+func (p *matrixSwapper) EvictionFor(target string, running []string) []string {
+	return p.solver.Solve(target, running).Evict
 }
 
-func (p *matrixPlanner) OnSwapStart(target string) {
-	running := p.runningModels()
+func (p *matrixSwapper) OnSwapStart(target string, running []string) {
 	result := p.solver.Solve(target, running)
 	switch {
 	case len(result.Evict) > 0:
@@ -69,32 +70,3 @@ func (p *matrixPlanner) OnSwapStart(target string) {
 		p.logger.Debugf("matrix: model=%s already running in set=%s dsl=%q", target, result.SetName, result.DSL)
 	}
 }
-
-func (p *matrixPlanner) runningModels() []string {
-	return p.runningSet(nil)
-}
-
-// runningSet returns the union of live processes (State != Stopped/Shutdown)
-// and any extra IDs the baseRouter has already committed to loading but which
-// the process state machine has not yet reflected.
-func (p *matrixPlanner) runningSet(alsoRunning []string) []string {
-	seen := make(map[string]struct{}, len(p.processes))
-	var running []string
-	for id, proc := range p.processes {
-		st := proc.State()
-		if st == process.StateStopped || st == process.StateShutdown {
-			continue
-		}
-		seen[id] = struct{}{}
-		running = append(running, id)
-	}
-	for _, id := range alsoRunning {
-		if _, dup := seen[id]; dup {
-			continue
-		}
-		seen[id] = struct{}{}
-		running = append(running, id)
-	}
-	sort.Strings(running)
-	return running
-}

internal/router/matrix_test.go

@@ -17,12 +17,14 @@ import (
 func newTestMatrix(t *testing.T, conf config.Config, expanded []config.ExpandedSet, evictCosts map[string]int, processes map[string]process.Process) *Matrix {
 	t.Helper()
 	logger := logmon.NewWriter(io.Discard)
-	planner := &matrixPlanner{
-		solver:    newMatrixSolver(expanded, evictCosts),
-		processes: processes,
-		logger:    logger,
+	swapper := &matrixSwapper{
+		solver: newMatrixSolver(expanded, evictCosts),
+		logger: logger,
+	}
+	base, err := newBaseRouter("matrix", conf, processes, logger, swapper)
+	if err != nil {
+		t.Fatalf("newBaseRouter: %v", err)
 	}
-	base := newBaseRouter("matrix", conf, processes, planner, logger)
 	base.testProcessed = make(chan struct{}, 64)
 	r := &Matrix{baseRouter: base}
 	go base.run()
@@ -153,8 +155,8 @@ func TestMatrix_CoexistingSetParallel(t *testing.T) {
 
 // TestMatrix_IncompatibleQueues verifies that the second request for a model
 // that cannot coexist with the in-flight first model queues until the first
-// completes, and then evicts it. This exercises the alsoRunning hint via the
-// matrix solver's union into runningSet.
+// completes, and then evicts it. This exercises the scheduler folding in-flight
+// swap targets into the running set it hands the swapper.
 func TestMatrix_IncompatibleQueues(t *testing.T) {
 	a := newFakeProcess("a")
 	pb := newFakeProcess("b")
@@ -173,8 +175,9 @@ func TestMatrix_IncompatibleQueues(t *testing.T) {
 	}()
 	waitProcessed(t, r.testProcessed, 1)
 
-	// B arrives before A transitions to StateStarting. The solver sees A via
-	// alsoRunning and returns evict=[a], so collidesWith forces B to queue.
+	// B arrives before A transitions to StateStarting. The running set the
+	// scheduler builds includes A (an in-flight swap target), so the solver
+	// returns evict=[a] and collidesWith forces B to queue.
 	w2 := httptest.NewRecorder()
 	done2 := make(chan struct{})
 	go func() {

internal/router/peer.go

@@ -15,6 +15,7 @@ import (
 
 	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
 type peerMember struct {
@@ -62,13 +63,12 @@ func NewPeer(cfg config.Config, logger *logmon.Monitor) (*Peer, error) {
 			IdleConnTimeout:       time.Duration(peer.Timeouts.IdleConn) * time.Second,
 		}
 
-		reverseProxy := httputil.NewSingleHostReverseProxy(peer.ProxyURL)
-		reverseProxy.Transport = peerTransport
-
-		originalDirector := reverseProxy.Director
-		reverseProxy.Director = func(req *http.Request) {
-			originalDirector(req)
-			req.Host = req.URL.Host
+		reverseProxy := &httputil.ReverseProxy{
+			Transport: peerTransport,
+			Rewrite: func(r *httputil.ProxyRequest) {
+				r.SetURL(peer.ProxyURL)
+				r.Out.Host = r.Out.URL.Host
+			},
 		}
 
 		reverseProxy.ModifyResponse = func(resp *http.Response) error {
@@ -147,22 +147,22 @@ func (r *Peer) Shutdown(timeout time.Duration) error {
 
 func (r *Peer) ServeHTTP(w http.ResponseWriter, req *http.Request) {
 	if r.shuttingDown.Load() {
-		SendError(w, req, fmt.Errorf("peer proxy is shutting down"))
+		shared.SendError(w, req, fmt.Errorf("peer proxy is shutting down"))
 		return
 	}
 	r.inflight.Add(1)
 	defer r.inflight.Done()
 
-	data, err := FetchContext(req, r.cfg)
+	data, err := shared.FetchContext(req, r.cfg)
 	if err != nil {
-		SendError(w, req, err)
+		shared.SendError(w, req, err)
 		return
 	}
 
 	pp, found := r.peers[data.ModelID]
 	if !found {
 		r.logger.Warnf("peer model not found: %s", data.ModelID)
-		SendError(w, req, ErrNoPeerModelFound)
+		shared.SendError(w, req, ErrNoPeerModelFound)
 		return
 	}
 

internal/router/peer_test.go

@@ -12,6 +12,7 @@ import (
 
 	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
 var testLogger = logmon.NewWriter(os.Stdout)
@@ -142,7 +143,7 @@ func TestPeer_ServeHTTP_Success(t *testing.T) {
 	}
 
 	req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "test-model", ModelID: "test-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "test-model", ModelID: "test-model"}))
 	w := httptest.NewRecorder()
 
 	pr.ServeHTTP(w, req)
@@ -178,7 +179,7 @@ func TestPeer_ServeHTTP_PeerModelNotFound(t *testing.T) {
 	}
 
 	req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "nonexistent-model", ModelID: "nonexistent-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "nonexistent-model", ModelID: "nonexistent-model"}))
 	w := httptest.NewRecorder()
 
 	pr.ServeHTTP(w, req)
@@ -212,7 +213,7 @@ func TestPeer_ServeHTTP_ApiKeyInjection(t *testing.T) {
 	}
 
 	req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "test-model", ModelID: "test-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "test-model", ModelID: "test-model"}))
 	w := httptest.NewRecorder()
 
 	pr.ServeHTTP(w, req)
@@ -246,7 +247,7 @@ func TestPeer_ServeHTTP_NoApiKey(t *testing.T) {
 	}
 
 	req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "test-model", ModelID: "test-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "test-model", ModelID: "test-model"}))
 	w := httptest.NewRecorder()
 
 	pr.ServeHTTP(w, req)
@@ -279,7 +280,7 @@ func TestPeer_ServeHTTP_HostHeaderSet(t *testing.T) {
 	}
 
 	req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "test-model", ModelID: "test-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "test-model", ModelID: "test-model"}))
 	w := httptest.NewRecorder()
 
 	pr.ServeHTTP(w, req)
@@ -311,7 +312,7 @@ func TestPeer_ServeHTTP_SSEHeaderModification(t *testing.T) {
 	}
 
 	req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "test-model", ModelID: "test-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "test-model", ModelID: "test-model"}))
 	w := httptest.NewRecorder()
 
 	pr.ServeHTTP(w, req)
@@ -347,7 +348,7 @@ func TestPeer_ServeHTTP_ShutdownRejectsNewRequests(t *testing.T) {
 	}
 
 	req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "test-model", ModelID: "test-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "test-model", ModelID: "test-model"}))
 	w := httptest.NewRecorder()
 
 	pr.ServeHTTP(w, req)
@@ -385,7 +386,7 @@ func TestPeer_ServeHTTP_WaitsForInflightDuringShutdown(t *testing.T) {
 	}
 
 	req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "test-model", ModelID: "test-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "test-model", ModelID: "test-model"}))
 
 	var wg sync.WaitGroup
 	wg.Add(1)
@@ -448,7 +449,7 @@ func TestPeer_ServeHTTP_ShutdownTimeoutCancelsInflight(t *testing.T) {
 	}
 
 	req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "test-model", ModelID: "test-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "test-model", ModelID: "test-model"}))
 
 	var wg sync.WaitGroup
 	wg.Add(1)
@@ -551,7 +552,7 @@ func TestPeer_ServeHTTP_ContextOverridesBodyModel(t *testing.T) {
 	body := strings.NewReader(`{"model":"body-model","prompt":"hello"}`)
 	req := httptest.NewRequest("POST", "/v1/chat/completions", body)
 	req.Header.Set("Content-Type", "application/json")
-	*req = *req.WithContext(SetContext(req.Context(), ReqContextData{Model: "context-model", ModelID: "context-model"}))
+	*req = *req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: "context-model", ModelID: "context-model"}))
 	w := httptest.NewRecorder()
 
 	pr.ServeHTTP(w, req)

internal/router/router.go

@@ -1,39 +1,18 @@
 package router
 
 import (
-	"bytes"
-	"context"
-	"errors"
-	"fmt"
-	"io"
 	"net/http"
-	"strings"
 	"time"
 
-	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
 	"github.com/mostlygeek/llama-swap/internal/process"
-	"github.com/tidwall/gjson"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
-type contextkey struct {
-	name string
-}
-
-type ReqContextData struct {
-	Model            string
-	ModelID          string
-	Streaming        bool
-	SendLoadingState bool
-}
-
 var (
-	ErrNoModelInContext  = fmt.Errorf("no model in request context")
-	ErrNoRouterFound     = fmt.Errorf("no router found for model")
-	ErrNoPeerModelFound  = fmt.Errorf("peer model not found")
-	ErrNoLocalModelFound = fmt.Errorf("local model not found")
-
-	ContextKey = &contextkey{"context"}
+	ErrNoRouterFound     = shared.ErrNoRouterFound
+	ErrNoPeerModelFound  = shared.ErrNoPeerModelFound
+	ErrNoLocalModelFound = shared.ErrNoLocalModelFound
 )
 
 type Router interface {
@@ -71,129 +50,3 @@ type LocalRouter interface {
 	// model is not known to this router.
 	ProcessLogger(modelID string) (*logmon.Monitor, bool)
 }
-
-// FetchContext will attempt to get the model id from the context then
-// from the model body. If it extracts the model from the body it will
-// store the model in the context for downstream handlers. An error
-// will be returned when model can not be fetch from either location.
-func FetchContext(r *http.Request, cfg config.Config) (ReqContextData, error) {
-	data, ok := ReadContext(r.Context())
-	if ok {
-		return data, nil
-	}
-
-	if data, err := ExtractContext(r); err == nil {
-		realName, _ := cfg.RealModelName(data.Model)
-		if realName == "" {
-			realName = data.Model
-		}
-		data.ModelID = realName
-		if mc, ok := cfg.Models[realName]; ok {
-			data.SendLoadingState = mc.SendLoadingState != nil && *mc.SendLoadingState
-		}
-		*r = *r.WithContext(SetContext(r.Context(), data))
-		return data, nil
-	}
-
-	return ReqContextData{}, ErrNoModelInContext
-}
-
-func SetContext(ctx context.Context, data ReqContextData) context.Context {
-	return context.WithValue(ctx, ContextKey, data)
-}
-
-func ReadContext(ctx context.Context) (ReqContextData, bool) {
-	data, ok := ctx.Value(ContextKey).(ReqContextData)
-	return data, ok
-}
-
-// ExtractContext pulls the model name from an HTTP request without consuming the
-// body. For GET requests it reads the "model" query parameter. For POST
-// requests it inspects Content-Type and parses JSON, multipart/form-data, or
-// application/x-www-form-urlencoded bodies. The request body is always restored
-// before returning so downstream handlers — including reverse proxies that
-// forward raw bytes upstream — can still read it.
-func ExtractContext(r *http.Request) (ReqContextData, error) {
-	if r.Method == http.MethodGet {
-		if model := r.URL.Query().Get("model"); model != "" {
-			return ReqContextData{Model: model, Streaming: r.URL.Query().Get("stream") == "true"}, nil
-		}
-		return ReqContextData{}, fmt.Errorf("missing 'model' query parameter")
-	}
-
-	bodyBytes, err := io.ReadAll(r.Body)
-	if err != nil {
-		return ReqContextData{}, fmt.Errorf("error reading request body: %w", err)
-	}
-	defer func() {
-		r.Body = io.NopCloser(bytes.NewReader(bodyBytes))
-	}()
-
-	contentType := r.Header.Get("Content-Type")
-
-	if strings.Contains(contentType, "application/json") {
-		model := gjson.GetBytes(bodyBytes, "model").String()
-		if model == "" {
-			return ReqContextData{}, fmt.Errorf("missing or empty 'model' in JSON body")
-		}
-		return ReqContextData{Model: model, Streaming: gjson.GetBytes(bodyBytes, "stream").Bool()}, nil
-	}
-
-	// Form parsers read from r.Body, so feed them a fresh reader over the
-	// buffered bytes. The deferred restore above will reset r.Body again
-	// after parsing.
-	r.Body = io.NopCloser(bytes.NewReader(bodyBytes))
-	if strings.Contains(contentType, "multipart/form-data") {
-		if err := r.ParseMultipartForm(32 << 20); err != nil {
-			return ReqContextData{}, fmt.Errorf("error parsing multipart form: %w", err)
-		}
-	} else {
-		if err := r.ParseForm(); err != nil {
-			return ReqContextData{}, fmt.Errorf("error parsing form: %w", err)
-		}
-	}
-
-	if model := r.FormValue("model"); model != "" {
-		return ReqContextData{Model: model, Streaming: r.FormValue("stream") == "true"}, nil
-	}
-
-	return ReqContextData{}, fmt.Errorf("missing 'model' parameter")
-}
-
-func SendError(w http.ResponseWriter, r *http.Request, err error) {
-	switch {
-	case errors.Is(err, ErrNoModelInContext):
-		SendResponse(w, r, http.StatusNotFound, "no model id could be identified")
-	case errors.Is(err, ErrNoPeerModelFound):
-		SendResponse(w, r, http.StatusNotFound, "no peer found for requested model")
-	case errors.Is(err, ErrNoLocalModelFound):
-		SendResponse(w, r, http.StatusNotFound, "no local server found for requested model")
-	case errors.Is(err, ErrNoRouterFound):
-		SendResponse(w, r, http.StatusNotFound, "no router for requested model")
-	default:
-		SendResponse(w, r, http.StatusInternalServerError, fmt.Sprintf("unspecific error: %v", err))
-	}
-}
-
-// SendResponse detects what content type the client prefers and returns an error response in that format.
-func SendResponse(w http.ResponseWriter, r *http.Request, status int, message string) {
-	// Check Accept header for preferred response format
-	acceptHeader := r.Header.Get("Accept")
-	if strings.Contains(acceptHeader, "text/plain") {
-		w.Header().Set("Content-Type", "text/plain")
-		w.WriteHeader(status)
-		w.Write([]byte(fmt.Sprintf("llama-swap: %s", message)))
-		return
-	}
-
-	if strings.Contains(acceptHeader, "text/html") {
-		w.Header().Set("Content-Type", "text/html")
-		w.WriteHeader(status)
-		w.Write([]byte(fmt.Sprintf(`<html><body><h1>llama-swap</h1><p>%s</p></body></html>`, message)))
-		return
-	}
-
-	w.Header().Set("Content-Type", "application/json")
-	w.WriteHeader(status)
-	w.Write([]byte(fmt.Sprintf(`{"src":"llama-swap", "error": "%s"}`, message)))
-}

internal/router/router_test.go

@@ -1,275 +0,0 @@
-package router
-
-import (
-	"bytes"
-	"context"
-	"io"
-	"mime/multipart"
-	"net/http"
-	"net/url"
-	"strings"
-	"testing"
-)
-
-func TestExtractContext_GET(t *testing.T) {
-	tests := []struct {
-		name      string
-		query     string
-		wantModel string
-		wantErr   bool
-	}{
-		{"model present", "model=llama3", "llama3", false},
-		{"model with slashes", "model=author/model-7b", "author/model-7b", false},
-		{"model missing", "", "", true},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			r, _ := http.NewRequest(http.MethodGet, "/?"+tt.query, nil)
-			got, err := ExtractContext(r)
-			if (err != nil) != tt.wantErr {
-				t.Fatalf("wantErr=%v got err=%v", tt.wantErr, err)
-			}
-			if got.Model != tt.wantModel {
-				t.Errorf("want %q got %q", tt.wantModel, got.Model)
-			}
-		})
-	}
-}
-
-func TestExtractContext_JSON(t *testing.T) {
-	tests := []struct {
-		name      string
-		body      string
-		wantModel string
-		wantErr   bool
-	}{
-		{"model present", `{"model":"llama3","stream":true}`, "llama3", false},
-		{"model with slashes", `{"model":"author/model-7b"}`, "author/model-7b", false},
-		{"model empty string", `{"model":""}`, "", true},
-		{"model key missing", `{"stream":true}`, "", true},
-		{"invalid json", `not-json`, "", true},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			r, _ := http.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(tt.body))
-			r.Header.Set("Content-Type", "application/json")
-			got, err := ExtractContext(r)
-			if (err != nil) != tt.wantErr {
-				t.Fatalf("wantErr=%v got err=%v", tt.wantErr, err)
-			}
-			if got.Model != tt.wantModel {
-				t.Errorf("want %q got %q", tt.wantModel, got.Model)
-			}
-		})
-	}
-}
-
-func TestExtractContext_URLEncodedForm(t *testing.T) {
-	tests := []struct {
-		name      string
-		formModel string
-		wantModel string
-		wantErr   bool
-	}{
-		{"model present", "whisper-1", "whisper-1", false},
-		{"model missing", "", "", true},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			form := url.Values{}
-			if tt.formModel != "" {
-				form.Set("model", tt.formModel)
-			}
-			r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", strings.NewReader(form.Encode()))
-			r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
-			got, err := ExtractContext(r)
-			if (err != nil) != tt.wantErr {
-				t.Fatalf("wantErr=%v got err=%v", tt.wantErr, err)
-			}
-			if got.Model != tt.wantModel {
-				t.Errorf("want %q got %q", tt.wantModel, got.Model)
-			}
-		})
-	}
-}
-
-func TestExtractContext_MultipartForm(t *testing.T) {
-	tests := []struct {
-		name      string
-		formModel string
-		wantModel string
-		wantErr   bool
-	}{
-		{"model present", "whisper-1", "whisper-1", false},
-		{"model missing", "", "", true},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			var buf bytes.Buffer
-			mw := multipart.NewWriter(&buf)
-			if tt.formModel != "" {
-				fw, _ := mw.CreateFormField("model")
-				fw.Write([]byte(tt.formModel))
-			}
-			mw.Close()
-
-			r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", &buf)
-			r.Header.Set("Content-Type", mw.FormDataContentType())
-			got, err := ExtractContext(r)
-			if (err != nil) != tt.wantErr {
-				t.Fatalf("wantErr=%v got err=%v", tt.wantErr, err)
-			}
-			if got.Model != tt.wantModel {
-				t.Errorf("want %q got %q", tt.wantModel, got.Model)
-			}
-		})
-	}
-}
-
-func TestExtractContext_JSONBodyRestored(t *testing.T) {
-	body := `{"model":"llama3","stream":true}`
-	r, _ := http.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(body))
-	r.Header.Set("Content-Type", "application/json")
-
-	if _, err := ExtractContext(r); err != nil {
-		t.Fatalf("ExtractContext: %v", err)
-	}
-
-	remaining, err := io.ReadAll(r.Body)
-	if err != nil {
-		t.Fatalf("reading body after ExtractContext: %v", err)
-	}
-	if string(remaining) != body {
-		t.Errorf("body not restored: want %q got %q", body, string(remaining))
-	}
-}
-
-func TestExtractContext_MultipartBodyRestored(t *testing.T) {
-	var buf bytes.Buffer
-	mw := multipart.NewWriter(&buf)
-	fw, _ := mw.CreateFormField("model")
-	fw.Write([]byte("whisper-1"))
-	ff, _ := mw.CreateFormFile("file", "audio.wav")
-	ff.Write([]byte("fake-audio-bytes"))
-	mw.Close()
-
-	original := buf.Bytes()
-
-	r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", bytes.NewReader(original))
-	r.Header.Set("Content-Type", mw.FormDataContentType())
-
-	if _, err := ExtractContext(r); err != nil {
-		t.Fatalf("ExtractContext: %v", err)
-	}
-
-	remaining, err := io.ReadAll(r.Body)
-	if err != nil {
-		t.Fatalf("reading body after ExtractContext: %v", err)
-	}
-	if !bytes.Equal(remaining, original) {
-		t.Errorf("multipart body not restored: want %d bytes got %d bytes", len(original), len(remaining))
-	}
-}
-
-func TestExtractContext_URLEncodedBodyRestored(t *testing.T) {
-	body := "model=whisper-1&extra=value"
-	r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", strings.NewReader(body))
-	r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
-
-	if _, err := ExtractContext(r); err != nil {
-		t.Fatalf("ExtractContext: %v", err)
-	}
-
-	remaining, err := io.ReadAll(r.Body)
-	if err != nil {
-		t.Fatalf("reading body after ExtractContext: %v", err)
-	}
-	if string(remaining) != body {
-		t.Errorf("url-encoded body not restored: want %q got %q", body, string(remaining))
-	}
-}
-
-func TestSetContext(t *testing.T) {
-	ctx := SetContext(context.Background(), ReqContextData{Model: "llama3", ModelID: "llama3"})
-	data, ok := ctx.Value(ContextKey).(ReqContextData)
-	if !ok {
-		t.Fatalf("ContextKey not set or wrong type")
-	}
-	if data.Model != "llama3" {
-		t.Errorf("want %q got %q", "llama3", data.Model)
-	}
-	if data.ModelID != "llama3" {
-		t.Errorf("want %q got %q", "llama3", data.ModelID)
-	}
-}
-
-func TestSetContext_WithAlias(t *testing.T) {
-	ctx := SetContext(context.Background(), ReqContextData{Model: "llama", ModelID: "llama3"})
-	data, _ := ctx.Value(ContextKey).(ReqContextData)
-	if data.Model != "llama" {
-		t.Errorf("want requested %q got %q", "llama", data.Model)
-	}
-	if data.ModelID != "llama3" {
-		t.Errorf("want real %q got %q", "llama3", data.ModelID)
-	}
-}
-
-func TestSetContext_DoesNotMutateParent(t *testing.T) {
-	parent := context.Background()
-	_ = SetContext(parent, ReqContextData{Model: "llama3", ModelID: "llama3"})
-	if v := parent.Value(ContextKey); v != nil {
-		t.Errorf("parent context was mutated: %v", v)
-	}
-}
-
-func TestReadContext(t *testing.T) {
-	tests := []struct {
-		name     string
-		ctx      context.Context
-		wantReq  string
-		wantReal string
-		wantBool bool
-	}{
-		{
-			name:     "model present, same name",
-			ctx:      SetContext(context.Background(), ReqContextData{Model: "llama3", ModelID: "llama3"}),
-			wantReq:  "llama3",
-			wantReal: "llama3",
-			wantBool: true,
-		},
-		{
-			name:     "model present, aliased",
-			ctx:      SetContext(context.Background(), ReqContextData{Model: "llama", ModelID: "llama3"}),
-			wantReq:  "llama",
-			wantReal: "llama3",
-			wantBool: true,
-		},
-		{
-			name:     "model absent",
-			ctx:      context.Background(),
-			wantReq:  "",
-			wantReal: "",
-			wantBool: false,
-		},
-		{
-			name:     "model is empty string",
-			ctx:      SetContext(context.Background(), ReqContextData{Model: "", ModelID: ""}),
-			wantReq:  "",
-			wantReal: "",
-			wantBool: true,
-		},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			gotData, ok := ReadContext(tt.ctx)
-			if gotData.Model != tt.wantReq || gotData.ModelID != tt.wantReal || ok != tt.wantBool {
-				t.Errorf("want (%q, %q, %v) got (%q, %q, %v)", tt.wantReq, tt.wantReal, tt.wantBool, gotData.Model, gotData.ModelID, ok)
-			}
-		})
-	}
-}

internal/router/scheduler/fifo.go

@@ -0,0 +1,489 @@
+package scheduler
+
+import (
+	"fmt"
+	"sort"
+	"strconv"
+	"time"
+
+	"github.com/mostlygeek/llama-swap/internal/config"
+	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"github.com/mostlygeek/llama-swap/internal/process"
+	"github.com/mostlygeek/llama-swap/internal/shared"
+)
+
+// defaultConcurrencyLimit caps simultaneous in-flight requests per model when
+// the model config leaves concurrencyLimit unset.
+const defaultConcurrencyLimit = 10
+
+// activeSwap tracks one in-flight swap and the callers waiting on it.
+type activeSwap struct {
+	modelID string
+	evict   []string
+	waiters []HandlerReq
+}
+
+// FIFO is the default scheduler. Requests are handled in a first-in, first-out order.
+// To reduce swapping requests for a model that is already running will be handled
+// immediately by the running process.
+//
+// Requests into this schedule are handled like this:
+//
+// A B C A B C --> A A B B C C
+//
+// The strategy is simple and reduces the number of swaps required.
+type FIFO struct {
+	name    string
+	logger  *logmon.Monitor
+	planner Swapper
+	cfg     config.FifoConfig
+	effects Effects
+
+	limits   map[string]int
+	active   map[string]*activeSwap
+	inFlight map[string]int
+	queued   []HandlerReq
+}
+
+// NewFIFO builds a FIFO scheduler. Per-model concurrency limits are derived
+// from models: each model's ConcurrencyLimit overrides defaultConcurrencyLimit
+// when set to a value greater than zero.
+func NewFIFO(name string, logger *logmon.Monitor, planner Swapper, cfg config.FifoConfig, models map[string]config.ModelConfig, eff Effects) *FIFO {
+	limits := make(map[string]int, len(models))
+	for id, mc := range models {
+		limit := defaultConcurrencyLimit
+		if mc.ConcurrencyLimit > 0 {
+			limit = mc.ConcurrencyLimit
+		}
+		limits[id] = limit
+	}
+
+	return &FIFO{
+		name:     name,
+		logger:   logger,
+		planner:  planner,
+		cfg:      cfg,
+		effects:  eff,
+		limits:   limits,
+		active:   make(map[string]*activeSwap),
+		inFlight: make(map[string]int),
+	}
+}
+
+// OnRequest decides what to do with one incoming ServeHTTP request. It never
+// blocks indefinitely: any work that has to wait (starting a process, stopping
+// siblings, waiting for ready) is deferred to a swap goroutine and reported back
+// via OnSwapDone.
+//
+// The decision tree, in order:
+//
+//  1. Unknown model — respond with ErrModelNotFound and move on.
+//  2. A swap to the same model is already in flight — attach this waiter so
+//     one swap serves all callers that asked for the same model.
+//  3. Fast path — the target process is already ready, the planner sees
+//     nothing to evict, and no in-flight swap is evicting it. Hand back its
+//     ServeHTTP immediately.
+//  4. Would collide with an in-flight swap (we'd stop their target, or they're
+//     stopping us) — park in the queue for OnSwapDone to drain.
+//  5. Would evict a process that is still handling requests — park in the
+//     queue. OnServeDone will retry when the busy process drains.
+//  6. Otherwise — start a new swap. This may run in parallel with other active
+//     swaps when their evict sets don't intersect.
+func (s *FIFO) OnRequest(req HandlerReq) {
+	// (1) Unknown model.
+	state, ok := s.effects.ModelState(req.Model)
+	if !ok {
+		s.logger.Debugf("%s: model %s not handled by this router", s.name, req.Model)
+		s.effects.GrantError(req, ErrModelNotFound)
+		return
+	}
+
+	// (2) Join an in-flight swap for the same model.
+	if sw, ok := s.active[req.Model]; ok {
+		s.logger.Debugf("%s: joining in-flight swap for model %s (%d waiters)", s.name, req.Model, len(sw.waiters)+1)
+		sw.waiters = append(sw.waiters, req)
+		return
+	}
+
+	running := s.runningSet(req.Model)
+	evict := s.planner.EvictionFor(req.Model, running)
+
+	// (3) Fast path: ready, nothing to evict, and nobody is evicting us.
+	if state == process.StateReady && len(evict) == 0 && !collidesWith(req.Model, evict, s.active) {
+		s.logger.Debugf("%s: fast-path serving model %s (already ready)", s.name, req.Model)
+		s.grantHandler(req, req.Model)
+		return
+	}
+
+	// (4) Collision with an in-flight swap — queue.
+	if collidesWith(req.Model, evict, s.active) {
+		s.logger.Debugf("%s: queuing request for model %s (collides with in-flight swap)", s.name, req.Model)
+		s.enqueue(req)
+		return
+	}
+
+	// (5) Would evict a busy process — queue until it drains.
+	if conflictsWithInFlight(evict, s.inFlight) {
+		s.logger.Debugf("%s: queuing request for model %s (would evict in-flight process)", s.name, req.Model)
+		s.enqueue(req)
+		return
+	}
+
+	// (6) Start a new (possibly parallel) swap.
+	s.logger.Debugf("%s: starting swap for model %s, evicting %v", s.name, req.Model, evict)
+	s.startSwap(req, evict, running)
+}
+
+// OnCancel removes a request whose client has disconnected from the queue and
+// from every in-flight swap's waiters. If the request was the sole waiter of an
+// active swap, the swap goroutine is left to complete on its own — OnSwapDone
+// will find no waiters and simply clean up. This prevents drainQueue from ever
+// starting a model load for a caller that is no longer there.
+func (s *FIFO) OnCancel(req HandlerReq) {
+	removed := false
+
+	// Prune from the queue.
+	if len(s.queued) > 0 {
+		kept := s.queued[:0]
+		for _, q := range s.queued {
+			if q.Respond == req.Respond {
+				removed = true
+				continue
+			}
+			kept = append(kept, q)
+		}
+		s.queued = kept
+	}
+
+	// Prune from any active swap's waiters.
+	for _, sw := range s.active {
+		filtered := sw.waiters[:0]
+		for _, w := range sw.waiters {
+			if w.Respond == req.Respond {
+				removed = true
+				continue
+			}
+			filtered = append(filtered, w)
+		}
+		sw.waiters = filtered
+	}
+
+	if removed {
+		s.logger.Debugf("%s: cancelled request for model %s pruned from scheduler", s.name, req.Model)
+		broadcastQueuePositions(s.queued)
+	}
+}
+
+// OnSwapDone fans the result out to every waiter that joined this swap, removes
+// the swap from the active map, then walks the queue once, promoting any items
+// that no longer collide with the remaining active set. FIFO order is preserved:
+// items still blocked stay in place.
+func (s *FIFO) OnSwapDone(ev SwapDone) {
+	sw, ok := s.active[ev.ModelID]
+	if !ok {
+		return
+	}
+	delete(s.active, ev.ModelID)
+
+	for _, w := range sw.waiters {
+		if ev.Err != nil {
+			s.effects.GrantError(w, ev.Err)
+		} else {
+			s.grantHandler(w, ev.ModelID)
+		}
+	}
+
+	s.drainQueue()
+}
+
+// OnServeDone decrements the per-model in-flight count and, when that drops to
+// zero, retries the queue: requests whose swap was deferred because they would
+// have evicted this (now-idle) process can now proceed.
+func (s *FIFO) OnServeDone(ev ServeDoneEvent) {
+	s.inFlight[ev.ModelID]--
+	if s.inFlight[ev.ModelID] <= 0 {
+		delete(s.inFlight, ev.ModelID)
+		s.drainQueue()
+	}
+}
+
+// OnUnload reconciles router-owned state with the impending Stop, performs the
+// Stop (synchronously, via Effects) so callers of Unload remain blocked until
+// each targeted process has exited, then drains the queue.
+func (s *FIFO) OnUnload(targets []string, timeout time.Duration) {
+	unloadErr := fmt.Errorf("%s: model unloaded", s.name)
+
+	targetSet := make(map[string]bool, len(targets))
+	for _, id := range targets {
+		targetSet[id] = true
+	}
+
+	// Release waiters of any in-flight swap whose target is being unloaded.
+	// The swap goroutine itself is left to finish on its own; when its
+	// SwapDone arrives, OnSwapDone will find no entry in active and drop it.
+	for id := range targetSet {
+		sw, ok := s.active[id]
+		if !ok {
+			continue
+		}
+		for _, w := range sw.waiters {
+			s.effects.GrantError(w, unloadErr)
+		}
+		delete(s.active, id)
+	}
+
+	// Drop queued requests addressed to unloaded models. Requests for other
+	// models stay queued and may benefit from drainQueue at the end.
+	if len(s.queued) > 0 {
+		kept := s.queued[:0]
+		for _, w := range s.queued {
+			if targetSet[w.Model] {
+				s.effects.GrantError(w, unloadErr)
+				continue
+			}
+			kept = append(kept, w)
+		}
+		s.queued = kept
+	}
+
+	// Stop the targeted processes. Done synchronously so Unload's caller can
+	// rely on "after Unload returns, the process is stopped". inFlight is
+	// intentionally NOT cleared here: each dying handler will fire its tracked
+	// serve and reach OnServeDone in the normal way.
+	s.effects.StopProcesses(timeout, targets)
+
+	// Removing entries from active above may have unblocked queued requests
+	// that previously collided with the now-cancelled swaps.
+	s.drainQueue()
+}
+
+// OnShutdown grants err to every waiter still held by the scheduler.
+func (s *FIFO) OnShutdown(err error) {
+	for _, sw := range s.active {
+		for _, w := range sw.waiters {
+			s.effects.GrantError(w, err)
+		}
+	}
+	for _, w := range s.queued {
+		s.effects.GrantError(w, err)
+	}
+}
+
+// grantHandler hands the caller a tracked handler for modelID and, only if the
+// caller was still there to receive it, bumps the in-flight count. Incrementing
+// when the grant failed would strand the counter and block future evictions.
+// Requests that would exceed the model's concurrency limit are rejected with a
+// shared.NewConcurrencyLimitError (HTTP 429 with Retry-After).
+func (s *FIFO) grantHandler(req HandlerReq, modelID string) {
+	if s.inFlight[modelID] >= s.limit(modelID) {
+		s.effects.GrantError(req, shared.ConcurrencyLimitError{})
+		return
+	}
+
+	if err := shared.SetReqData(req.Ctx, "fifo_priority", strconv.Itoa(s.cfg.Priority[req.Model])); err != nil {
+		s.logger.Debugf("failed to set fifo_priority metadata: %v", err)
+	}
+
+	if s.effects.GrantServe(req, modelID) {
+		s.inFlight[modelID]++
+	}
+}
+
+// limit returns the per-model concurrency cap, defaulting to
+// defaultConcurrencyLimit when the model has no explicit entry.
+func (s *FIFO) limit(modelID string) int {
+	if l, ok := s.limits[modelID]; ok {
+		return l
+	}
+	return defaultConcurrencyLimit
+}
+
+// startSwap records the swap as active and launches it via Effects. running is
+// the set EvictionFor saw, forwarded to OnSwapStart so the planner logs against
+// the same picture it decided on.
+func (s *FIFO) startSwap(initial HandlerReq, evict, running []string) {
+	s.active[initial.Model] = &activeSwap{
+		modelID: initial.Model,
+		evict:   evict,
+		waiters: []HandlerReq{initial},
+	}
+	s.planner.OnSwapStart(initial.Model, running)
+	s.effects.StartSwap(initial.Model, evict)
+}
+
+// enqueue inserts req into the queue in priority order: it goes just before the
+// first queued item whose priority is strictly lower, so higher-priority models
+// are serviced first while equal-priority requests keep their arrival (FIFO)
+// order. Priorities come from the FifoConfig; unlisted models default to 0.
+func (s *FIFO) enqueue(req HandlerReq) {
+	p := s.cfg.Priority[req.Model]
+	i := len(s.queued)
+	for j, q := range s.queued {
+		if s.cfg.Priority[q.Model] < p {
+			i = j
+			break
+		}
+	}
+	s.queued = append(s.queued, HandlerReq{})
+	copy(s.queued[i+1:], s.queued[i:])
+	s.queued[i] = req
+	broadcastQueuePositions(s.queued)
+}
+
+// drainQueue walks the queued requests in order, re-running the OnRequest
+// decision tree against the (now smaller) active set. Items that can now start
+// or join become satisfied; items still blocked remain queued in original order
+// so they get another chance on the next swap completion.
+func (s *FIFO) drainQueue() {
+	if len(s.queued) == 0 {
+		return
+	}
+	pending := s.queued
+	var remaining []HandlerReq
+	for _, req := range pending {
+		state, ok := s.effects.ModelState(req.Model)
+		if !ok {
+			s.effects.GrantError(req, ErrModelNotFound)
+			continue
+		}
+		if sw, ok := s.active[req.Model]; ok {
+			s.logger.Debugf("%s: queued request for model %s now joining in-flight swap", s.name, req.Model)
+			sw.waiters = append(sw.waiters, req)
+			continue
+		}
+		running := s.runningSet(req.Model)
+		evict := s.planner.EvictionFor(req.Model, running)
+		if state == process.StateReady && len(evict) == 0 && !collidesWith(req.Model, evict, s.active) {
+			s.logger.Debugf("%s: queued request for model %s now served fast-path", s.name, req.Model)
+			s.grantHandler(req, req.Model)
+			continue
+		}
+		if collidesWith(req.Model, evict, s.active) {
+			remaining = append(remaining, req)
+			continue
+		}
+		if conflictsWithInFlight(evict, s.inFlight) {
+			remaining = append(remaining, req)
+			continue
+		}
+		s.logger.Debugf("%s: queued request for model %s now starting swap, evicting %v", s.name, req.Model, evict)
+		s.startSwap(req, evict, running)
+	}
+	s.queued = remaining
+	broadcastQueuePositions(s.queued)
+}
+
+// runningSet is the live model set handed to the Swapper: every process the
+// baseRouter reports as running, unioned with the targets of in-flight swaps
+// (excluding excludeActive, the model whose own swap is being decided — its
+// in-flight entry must not count as "already running"). The result is sorted so
+// eviction decisions derived from it are deterministic.
+func (s *FIFO) runningSet(excludeActive string) []string {
+	seen := make(map[string]struct{})
+	var out []string
+	add := func(id string) {
+		if _, dup := seen[id]; dup {
+			return
+		}
+		seen[id] = struct{}{}
+		out = append(out, id)
+	}
+	for id := range s.effects.RunningModels() {
+		add(id)
+	}
+	for _, id := range activeTargets(s.active, excludeActive) {
+		add(id)
+	}
+	sort.Strings(out)
+	return out
+}
+
+// activeTargets returns the IDs of every in-flight swap target except exclude.
+// The planner uses this to account for models committed to but not yet reflected
+// in process state.
+func activeTargets(active map[string]*activeSwap, exclude string) []string {
+	if len(active) == 0 {
+		return nil
+	}
+	out := make([]string, 0, len(active))
+	for id := range active {
+		if id == exclude {
+			continue
+		}
+		out = append(out, id)
+	}
+	return out
+}
+
+// collidesWith reports whether a new swap with this target and evict set can
+// safely run alongside the currently active swaps. Same-target callers should
+// JOIN (handled before this) — they do not collide with themselves.
+func collidesWith(target string, evict []string, active map[string]*activeSwap) bool {
+	for id, sw := range active {
+		if id == target {
+			continue
+		}
+		if containsString(evict, id) {
+			return true
+		}
+		if containsString(sw.evict, target) {
+			return true
+		}
+		if slicesOverlap(evict, sw.evict) {
+			return true
+		}
+	}
+	return false
+}
+
+// slicesOverlap reports whether xs and ys share any common element.
+func slicesOverlap(xs, ys []string) bool {
+	for _, x := range xs {
+		if containsString(ys, x) {
+			return true
+		}
+	}
+	return false
+}
+
+// conflictsWithInFlight reports whether any model in evict is still handling
+// requests. Stopping a busy process would cancel its callers' connections, so
+// the scheduler defers the swap until those callers finish.
+func conflictsWithInFlight(evict []string, inFlight map[string]int) bool {
+	for _, m := range evict {
+		if inFlight[m] > 0 {
+			return true
+		}
+	}
+	return false
+}
+
+func containsString(xs []string, s string) bool {
+	for _, x := range xs {
+		if x == s {
+			return true
+		}
+	}
+	return false
+}
+
+// broadcastQueuePositions sends each queued request its current 1-indexed
+// position. Sends are non-blocking: if the channel is full, the old value is
+// drained first so the consumer always sees the latest position.
+func broadcastQueuePositions(queued []HandlerReq) {
+	for i, req := range queued {
+		pos := i + 1
+		select {
+		case req.PositionCh <- pos:
+		default:
+			select {
+			case <-req.PositionCh:
+			default:
+			}
+			select {
+			case req.PositionCh <- pos:
+			default:
+			}
+		}
+	}
+}

internal/router/scheduler/fifo_test.go

@@ -0,0 +1,779 @@
+package scheduler
+
+import (
+	"context"
+	"errors"
+	"io"
+	"net/http"
+	"testing"
+	"time"
+
+	"github.com/mostlygeek/llama-swap/internal/config"
+	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"github.com/mostlygeek/llama-swap/internal/process"
+	"github.com/mostlygeek/llama-swap/internal/shared"
+)
+
+// FIFO methods all run on the router's single run-loop goroutine, so these
+// tests drive them directly and synchronously. A swap is "completed" by calling
+// OnSwapDone, a served request "finishes" by calling OnServeDone — exactly the
+// events the run loop would deliver. fakeEffects records every side-effect and
+// stubPlanner supplies a fixed eviction set per target.
+
+// stubPlanner returns a fixed eviction list per target.
+type stubPlanner struct {
+	evict map[string][]string
+}
+
+func (s *stubPlanner) EvictionFor(target string, _ []string) []string {
+	if s.evict == nil {
+		return nil
+	}
+	return s.evict[target]
+}
+
+func (s *stubPlanner) OnSwapStart(string, []string) {}
+
+// grantRec is one GrantError / GrantServe call. err!=nil marks an error grant;
+// otherwise it is a serve grant and serve reports whether the caller received it.
+type grantRec struct {
+	model string
+	err   error
+	serve bool
+}
+
+type startRec struct {
+	model string
+	evict []string
+}
+
+type stopRec struct {
+	timeout time.Duration
+	ids     []string
+}
+
+// fakeEffects is an in-memory scheduler.Effects. Tests program process states
+// and GrantServe outcomes, then assert on the recorded calls.
+type fakeEffects struct {
+	states       map[string]process.ProcessState // model -> state; missing => not handled
+	serveResult  map[string]bool                 // GrantServe return per model (default true)
+	lastServeReq HandlerReq
+
+	starts []startRec
+	grants []grantRec
+	stops  []stopRec
+}
+
+func newFakeEffects() *fakeEffects {
+	return &fakeEffects{
+		states:      map[string]process.ProcessState{},
+		serveResult: map[string]bool{},
+	}
+}
+
+func (f *fakeEffects) ModelState(modelID string) (process.ProcessState, bool) {
+	st, ok := f.states[modelID]
+	return st, ok
+}
+
+func (f *fakeEffects) RunningModels() map[string]process.ProcessState {
+	out := make(map[string]process.ProcessState)
+	for id, st := range f.states {
+		if st == process.StateStopped || st == process.StateShutdown {
+			continue
+		}
+		out[id] = st
+	}
+	return out
+}
+
+func (f *fakeEffects) StartSwap(modelID string, evict []string) {
+	f.starts = append(f.starts, startRec{model: modelID, evict: evict})
+}
+
+func (f *fakeEffects) GrantError(req HandlerReq, err error) {
+	f.grants = append(f.grants, grantRec{model: req.Model, err: err})
+}
+
+func (f *fakeEffects) GrantServe(req HandlerReq, modelID string) bool {
+	ok := true
+	if v, set := f.serveResult[modelID]; set {
+		ok = v
+	}
+	f.lastServeReq = req
+	f.grants = append(f.grants, grantRec{model: modelID, serve: ok})
+	return ok
+}
+
+func (f *fakeEffects) StopProcesses(timeout time.Duration, ids []string) {
+	f.stops = append(f.stops, stopRec{timeout: timeout, ids: ids})
+}
+
+// served counts grants that handed modelID a handler and were received.
+func (f *fakeEffects) served(modelID string) int {
+	n := 0
+	for _, g := range f.grants {
+		if g.err == nil && g.serve && g.model == modelID {
+			n++
+		}
+	}
+	return n
+}
+
+// errored counts error grants, optionally filtered by model ("" = any).
+func (f *fakeEffects) errored(model string) int {
+	n := 0
+	for _, g := range f.grants {
+		if g.err != nil && (model == "" || g.model == model) {
+			n++
+		}
+	}
+	return n
+}
+
+// startsFor counts StartSwap calls for modelID.
+func (f *fakeEffects) startsFor(modelID string) int {
+	n := 0
+	for _, s := range f.starts {
+		if s.model == modelID {
+			n++
+		}
+	}
+	return n
+}
+
+func newFIFO(planner Swapper, eff Effects) *FIFO {
+	return NewFIFO("test", logmon.NewWriter(io.Discard), planner, config.FifoConfig{}, nil, eff)
+}
+
+func req(model string) HandlerReq { return HandlerReq{Model: model} }
+
+// reqCh creates a HandlerReq with a unique Respond channel so OnCancel can
+// identify it among queued requests and swap waiters.
+func reqCh(model string) HandlerReq {
+	return HandlerReq{
+		Model:   model,
+		Respond: make(chan HandlerResp, 1),
+	}
+}
+
+func TestFIFO_FastPath(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateReady
+	s := newFIFO(&stubPlanner{}, eff)
+
+	s.OnRequest(req("a"))
+
+	if got := eff.startsFor("a"); got != 0 {
+		t.Errorf("StartSwap calls=%d want 0 (fast path should not swap)", got)
+	}
+	if got := eff.served("a"); got != 1 {
+		t.Errorf("served(a)=%d want 1", got)
+	}
+}
+
+func TestFIFO_GrantSetsPriorityMetadata(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateReady
+	cfg := config.FifoConfig{Priority: map[string]int{"a": 7}}
+	s := NewFIFO("test", logmon.NewWriter(io.Discard), &stubPlanner{}, cfg, nil, eff)
+
+	ctx := shared.SetContext(context.Background(), shared.ReqContextData{ModelID: "a", Metadata: make(map[string]string)})
+	s.OnRequest(HandlerReq{Model: "a", Ctx: ctx})
+
+	if got := eff.served("a"); got != 1 {
+		t.Fatalf("served(a)=%d want 1", got)
+	}
+	data, ok := shared.ReadContext(eff.lastServeReq.Ctx)
+	if !ok {
+		t.Fatal("context data missing from granted request")
+	}
+	if data.Metadata["fifo_priority"] != "7" {
+		t.Errorf("fifo_priority = %q, want 7", data.Metadata["fifo_priority"])
+	}
+}
+
+func TestFIFO_ModelNotFound(t *testing.T) {
+	eff := newFakeEffects() // no states => model unknown
+	s := newFIFO(&stubPlanner{}, eff)
+
+	s.OnRequest(req("ghost"))
+
+	if got := len(eff.starts); got != 0 {
+		t.Errorf("StartSwap calls=%d want 0", got)
+	}
+	if eff.errored("ghost") != 1 {
+		t.Fatalf("want 1 error grant for ghost, grants=%+v", eff.grants)
+	}
+	if !errors.Is(eff.grants[0].err, ErrModelNotFound) {
+		t.Errorf("err=%v want ErrModelNotFound", eff.grants[0].err)
+	}
+}
+
+func TestFIFO_OnDemandStartThenServe(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	s := newFIFO(&stubPlanner{}, eff)
+
+	s.OnRequest(req("a"))
+	if got := eff.startsFor("a"); got != 1 {
+		t.Fatalf("StartSwap(a)=%d want 1", got)
+	}
+	if got := eff.served("a"); got != 0 {
+		t.Errorf("served(a)=%d want 0 before swap completes", got)
+	}
+
+	// Swap finishes, model is now ready.
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+
+	if got := eff.served("a"); got != 1 {
+		t.Errorf("served(a)=%d want 1 after swap done", got)
+	}
+}
+
+func TestFIFO_JoinInFlightSwap(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	s := newFIFO(&stubPlanner{}, eff)
+
+	s.OnRequest(req("a")) // starts swap
+	s.OnRequest(req("a")) // joins
+	s.OnRequest(req("a")) // joins
+
+	if got := eff.startsFor("a"); got != 1 {
+		t.Fatalf("StartSwap(a)=%d want 1 (all three share one swap)", got)
+	}
+
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+
+	if got := eff.served("a"); got != 3 {
+		t.Errorf("served(a)=%d want 3 (one swap serves all waiters)", got)
+	}
+}
+
+func TestFIFO_SwapDoneError_FailsAllWaiters(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	s := newFIFO(&stubPlanner{}, eff)
+
+	s.OnRequest(req("a"))
+	s.OnRequest(req("a"))
+
+	s.OnSwapDone(SwapDone{ModelID: "a", Err: errors.New("boom")})
+
+	if eff.served("a") != 0 {
+		t.Errorf("served(a)=%d want 0 on swap error", eff.served("a"))
+	}
+	if eff.errored("a") != 2 {
+		t.Errorf("errored(a)=%d want 2 (both waiters fail)", eff.errored("a"))
+	}
+}
+
+// TestFIFO_QueueOnEvictionCollision covers a request whose target evicts the
+// model currently being swapped: it must queue until that swap finishes AND its
+// served request drains, because starting it would stop a busy process.
+func TestFIFO_QueueOnEvictionCollision(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	// Loading b evicts a.
+	s := newFIFO(&stubPlanner{evict: map[string][]string{"b": {"a"}}}, eff)
+
+	s.OnRequest(req("a")) // StartSwap(a)
+	s.OnRequest(req("b")) // collides with a's in-flight swap -> queue
+	if got := eff.startsFor("b"); got != 0 {
+		t.Fatalf("b started early: StartSwap(b)=%d want 0", got)
+	}
+
+	// a becomes ready and is granted (now serving, inFlight[a]=1).
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+	if got := eff.startsFor("b"); got != 0 {
+		t.Fatalf("b started while a is serving: StartSwap(b)=%d want 0", got)
+	}
+
+	// a's request finishes -> a no longer in-flight -> b may now swap.
+	s.OnServeDone(ServeDoneEvent{ModelID: "a"})
+	if got := eff.startsFor("b"); got != 1 {
+		t.Fatalf("StartSwap(b)=%d want 1 after a drained", got)
+	}
+	if got := eff.starts[len(eff.starts)-1].evict; len(got) != 1 || got[0] != "a" {
+		t.Errorf("b swap evict=%v want [a]", got)
+	}
+}
+
+// TestFIFO_DisjointSwapsRunInParallel verifies two requests with
+// non-conflicting evict sets both start without waiting for each other.
+func TestFIFO_DisjointSwapsRunInParallel(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	s := newFIFO(&stubPlanner{}, eff) // empty evicts
+
+	s.OnRequest(req("a"))
+	s.OnRequest(req("b"))
+
+	if eff.startsFor("a") != 1 || eff.startsFor("b") != 1 {
+		t.Fatalf("StartSwap a=%d b=%d want 1 each (parallel)", eff.startsFor("a"), eff.startsFor("b"))
+	}
+}
+
+// TestFIFO_OverlappingEvictSetsDoNotRunInParallel verifies two swaps with
+// different targets that evict the *same* model do not run concurrently: the
+// second must queue rather than double-evict the shared model. Neither target is
+// in the other's evict set, so this is only caught by the evict-set overlap
+// check in collidesWith.
+func TestFIFO_OverlappingEvictSetsDoNotRunInParallel(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	eff.states["x"] = process.StateReady // shared eviction target, running
+	// Loading a or b both require evicting x.
+	s := newFIFO(&stubPlanner{evict: map[string][]string{"a": {"x"}, "b": {"x"}}}, eff)
+
+	s.OnRequest(req("a")) // StartSwap(a, [x])
+	s.OnRequest(req("b")) // overlaps a's evict set ([x]) -> queue
+	if eff.startsFor("a") != 1 {
+		t.Fatalf("StartSwap(a)=%d want 1", eff.startsFor("a"))
+	}
+	if got := eff.startsFor("b"); got != 0 {
+		t.Fatalf("b started in parallel while a evicts x: StartSwap(b)=%d want 0", got)
+	}
+
+	// a's swap completes and x is gone; b can now evict nothing and start.
+	eff.states["a"] = process.StateReady
+	eff.states["x"] = process.StateStopped
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+	if got := eff.startsFor("b"); got != 1 {
+		t.Fatalf("StartSwap(b)=%d want 1 after a's swap drained", got)
+	}
+}
+
+// TestFIFO_QueueDrainPromotesMultiple verifies completing one swap unblocks
+// every queued request that no longer collides — they all start together.
+func TestFIFO_QueueDrainPromotesMultiple(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	eff.states["c"] = process.StateStopped
+	// a's swap evicts both b and c; b and c evict nothing.
+	s := newFIFO(&stubPlanner{evict: map[string][]string{"a": {"b", "c"}}}, eff)
+
+	s.OnRequest(req("a")) // StartSwap(a, [b,c])
+	s.OnRequest(req("b")) // collides (in a's evict set) -> queue
+	s.OnRequest(req("c")) // collides -> queue
+	if eff.startsFor("b") != 0 || eff.startsFor("c") != 0 {
+		t.Fatalf("b/c started early")
+	}
+
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+
+	// b and c have empty evict sets and don't evict a, so both start now.
+	if eff.startsFor("b") != 1 || eff.startsFor("c") != 1 {
+		t.Fatalf("StartSwap b=%d c=%d want 1 each after a done", eff.startsFor("b"), eff.startsFor("c"))
+	}
+	if eff.served("a") != 1 {
+		t.Errorf("served(a)=%d want 1", eff.served("a"))
+	}
+}
+
+// TestFIFO_QueueCollation verifies duplicate requests collapse into one swap
+// per model: the second request for each model joins the active swap (at arrival
+// or at drain time) rather than triggering its own swap.
+func TestFIFO_QueueCollation(t *testing.T) {
+	eff := newFakeEffects()
+	for _, id := range []string{"a", "b", "c"} {
+		eff.states[id] = process.StateStopped
+	}
+	// Each model evicts the other two: all swaps are mutually exclusive.
+	s := newFIFO(&stubPlanner{evict: map[string][]string{
+		"a": {"b", "c"},
+		"b": {"a", "c"},
+		"c": {"a", "b"},
+	}}, eff)
+
+	for _, id := range []string{"a", "b", "c", "a", "b", "c"} {
+		s.OnRequest(req(id))
+	}
+
+	// Drain a, then its served requests, which promotes b; repeat for b -> c.
+	drain := func(model string, waiters int) {
+		eff.states[model] = process.StateReady
+		s.OnSwapDone(SwapDone{ModelID: model})
+		for i := 0; i < waiters; i++ {
+			s.OnServeDone(ServeDoneEvent{ModelID: model})
+		}
+	}
+	drain("a", 2)
+	drain("b", 2)
+	drain("c", 2)
+
+	for _, id := range []string{"a", "b", "c"} {
+		if got := eff.startsFor(id); got != 1 {
+			t.Errorf("StartSwap(%s)=%d want 1 (collation)", id, got)
+		}
+		if got := eff.served(id); got != 2 {
+			t.Errorf("served(%s)=%d want 2", id, got)
+		}
+	}
+}
+
+// TestFIFO_NoSwapWhileServing verifies a model still handling requests is not
+// evicted: the evicting request waits until every in-flight request drains.
+func TestFIFO_NoSwapWhileServing(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateReady
+	eff.states["b"] = process.StateStopped
+	s := newFIFO(&stubPlanner{evict: map[string][]string{"b": {"a"}}}, eff)
+
+	s.OnRequest(req("a")) // fast path, inFlight[a]=1
+	s.OnRequest(req("a")) // fast path, inFlight[a]=2
+	s.OnRequest(req("b")) // would evict busy a -> queue
+	if eff.startsFor("b") != 0 {
+		t.Fatalf("b started while a serving")
+	}
+
+	s.OnServeDone(ServeDoneEvent{ModelID: "a"}) // inFlight[a]=1
+	if eff.startsFor("b") != 0 {
+		t.Fatalf("b started while a still serving one request")
+	}
+
+	s.OnServeDone(ServeDoneEvent{ModelID: "a"}) // inFlight[a]=0
+	if eff.startsFor("b") != 1 {
+		t.Fatalf("StartSwap(b)=%d want 1 after a fully drained", eff.startsFor("b"))
+	}
+}
+
+// TestFIFO_GrantServeFalseDoesNotLeakInFlight verifies that when a caller has
+// walked away (GrantServe returns false) the in-flight count is not bumped, so a
+// later evicting request is not blocked forever.
+func TestFIFO_GrantServeFalseDoesNotLeakInFlight(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	eff.serveResult["a"] = false // a's waiter is gone by grant time
+	s := newFIFO(&stubPlanner{evict: map[string][]string{"b": {"a"}}}, eff)
+
+	s.OnRequest(req("a"))
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"}) // grant fails, inFlight[a] stays 0
+
+	// b evicts a; since a is not in-flight, b should start immediately.
+	s.OnRequest(req("b"))
+	if eff.startsFor("b") != 1 {
+		t.Fatalf("StartSwap(b)=%d want 1 (no leaked in-flight on a)", eff.startsFor("b"))
+	}
+}
+
+// TestFIFO_OnShutdown_FailsAllWaiters verifies shutdown errors every waiter the
+// scheduler holds: active-swap waiters and queued requests alike.
+func TestFIFO_OnShutdown_FailsAllWaiters(t *testing.T) {
+	eff := newFakeEffects()
+	for _, id := range []string{"a", "b", "c"} {
+		eff.states[id] = process.StateStopped
+	}
+	// a and b load in parallel; c collides with both and queues.
+	s := newFIFO(&stubPlanner{evict: map[string][]string{"c": {"a", "b"}}}, eff)
+
+	s.OnRequest(req("a")) // StartSwap(a)
+	s.OnRequest(req("a")) // join a
+	s.OnRequest(req("b")) // StartSwap(b)
+	s.OnRequest(req("b")) // join b
+	s.OnRequest(req("c")) // queued
+
+	s.OnShutdown(errors.New("shutting down"))
+
+	if got := eff.errored(""); got != 5 {
+		t.Errorf("error grants=%d want 5 (2 a + 2 b + 1 c)", got)
+	}
+}
+
+func TestFIFO_OnUnload_ReleasesActiveWaiters(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	s := newFIFO(&stubPlanner{}, eff)
+
+	s.OnRequest(req("a")) // active swap a with one waiter
+	s.OnRequest(req("a")) // join
+
+	s.OnUnload([]string{"a"}, time.Second)
+
+	if got := eff.errored("a"); got != 2 {
+		t.Errorf("errored(a)=%d want 2 (active swap waiters released)", got)
+	}
+	if len(eff.stops) != 1 || len(eff.stops[0].ids) != 1 || eff.stops[0].ids[0] != "a" {
+		t.Errorf("StopProcesses=%+v want one call stopping [a]", eff.stops)
+	}
+	if eff.stops[0].timeout != time.Second {
+		t.Errorf("StopProcesses timeout=%v want 1s", eff.stops[0].timeout)
+	}
+}
+
+func TestFIFO_OnUnload_DropsQueuedRequests(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	// b evicts a, so a request for b queues while a is loading.
+	s := newFIFO(&stubPlanner{evict: map[string][]string{"b": {"a"}}}, eff)
+
+	s.OnRequest(req("a")) // StartSwap(a)
+	s.OnRequest(req("b")) // queued
+
+	s.OnUnload([]string{"b"}, time.Second)
+
+	if got := eff.errored("b"); got != 1 {
+		t.Errorf("errored(b)=%d want 1 (queued request dropped)", got)
+	}
+	if got := eff.startsFor("b"); got != 0 {
+		t.Errorf("StartSwap(b)=%d want 0 (b should never start)", got)
+	}
+	// a's swap is untouched: its waiter is neither served nor errored yet.
+	if eff.served("a") != 0 || eff.errored("a") != 0 {
+		t.Errorf("a swap should be untouched: served=%d errored=%d", eff.served("a"), eff.errored("a"))
+	}
+}
+
+// TestFIFO_PriorityQueueOrder verifies queued requests are ordered by descending
+// priority, with arrival (FIFO) order preserved among equal-priority models.
+func TestFIFO_PriorityQueueOrder(t *testing.T) {
+	eff := newFakeEffects()
+	for _, m := range []string{"z", "A", "B", "C", "D"} {
+		eff.states[m] = process.StateStopped
+	}
+	// z's swap evicts every other model, so any request that arrives while z is
+	// loading collides with z's in-flight swap and parks in the queue.
+	planner := &stubPlanner{evict: map[string][]string{"z": {"A", "B", "C", "D"}}}
+	cfg := config.FifoConfig{Priority: map[string]int{"A": 10, "B": 5, "C": 5, "D": 1}}
+	s := NewFIFO("test", logmon.NewWriter(io.Discard), planner, cfg, nil, eff)
+
+	s.OnRequest(req("z")) // StartSwap(z, [A,B,C,D])
+
+	// Arrive out of priority order; B before C exercises FIFO tie-breaking.
+	for _, m := range []string{"B", "D", "C", "A"} {
+		s.OnRequest(req(m))
+	}
+
+	got := make([]string, len(s.queued))
+	for i, q := range s.queued {
+		got[i] = q.Model
+	}
+	want := []string{"A", "B", "C", "D"}
+	if len(got) != len(want) {
+		t.Fatalf("queue=%v want %v", got, want)
+	}
+	for i := range want {
+		if got[i] != want[i] {
+			t.Fatalf("queue=%v want %v", got, want)
+		}
+	}
+}
+
+// TestFIFO_OnCancel_QueuedRequest verifies that cancelling a queued request
+// prevents drainQueue from ever starting a model load for it. Without OnCancel
+// the dead request would sit in the queue until a drain triggers a wasted swap.
+func TestFIFO_OnCancel_QueuedRequest(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	// b evicts a, so a request for b queues while a is loading.
+	s := newFIFO(&stubPlanner{evict: map[string][]string{"b": {"a"}}}, eff)
+
+	s.OnRequest(req("a")) // StartSwap(a)
+
+	cancelledReq := reqCh("b")
+	s.OnRequest(cancelledReq) // queued (collides with a's in-flight swap)
+	if len(s.queued) != 1 {
+		t.Fatalf("queue len=%d want 1 before cancel", len(s.queued))
+	}
+
+	// Client disconnects.
+	s.OnCancel(cancelledReq)
+
+	if len(s.queued) != 0 {
+		t.Fatalf("queue len=%d want 0 after cancel", len(s.queued))
+	}
+
+	// a's swap finishes; drainQueue runs but b is gone — no swap for b.
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+
+	if got := eff.startsFor("b"); got != 0 {
+		t.Errorf("StartSwap(b)=%d want 0 (cancelled request should not trigger a load)", got)
+	}
+}
+
+// TestFIFO_OnCancel_SwapWaiter verifies that cancelling a request that joined an
+// in-flight swap removes it from the waiter list. When the swap completes, the
+// cancelled waiter receives no grant and does not bump the in-flight count.
+func TestFIFO_OnCancel_SwapWaiter(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	s := newFIFO(&stubPlanner{}, eff)
+
+	liveReq := reqCh("a")
+	cancelledReq := reqCh("a")
+	s.OnRequest(liveReq)      // starts swap
+	s.OnRequest(cancelledReq) // joins
+
+	if sw := s.active["a"]; len(sw.waiters) != 2 {
+		t.Fatalf("waiters=%d want 2", len(sw.waiters))
+	}
+
+	s.OnCancel(cancelledReq)
+
+	if sw := s.active["a"]; len(sw.waiters) != 1 {
+		t.Fatalf("waiters=%d want 1 after cancel", len(sw.waiters))
+	}
+
+	// Swap finishes: only the live waiter is granted.
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+
+	if got := eff.served("a"); got != 1 {
+		t.Errorf("served(a)=%d want 1 (only the non-cancelled waiter)", got)
+	}
+}
+
+// TestFIFO_OnCancel_NotPresent is a no-op: cancelling a request that was already
+// granted (and is no longer queued or waiting) must not affect anything.
+func TestFIFO_OnCancel_NotPresent(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateReady
+	s := newFIFO(&stubPlanner{}, eff)
+
+	r := reqCh("a")
+	s.OnRequest(r) // fast-path served immediately
+
+	// Cancel after grant — should be a harmless no-op.
+	s.OnCancel(r)
+
+	if got := eff.served("a"); got != 1 {
+		t.Errorf("served(a)=%d want 1 (cancel of granted request is a no-op)", got)
+	}
+	if len(s.queued) != 0 {
+		t.Errorf("queue should be empty, len=%d", len(s.queued))
+	}
+}
+
+// newFIFOWithLimit builds a FIFO whose single model has the given concurrency
+// limit, already in StateReady so every request exercises the fast path.
+func newFIFOWithLimit(t *testing.T, model string, limit int) (*FIFO, *fakeEffects) {
+	t.Helper()
+	eff := newFakeEffects()
+	eff.states[model] = process.StateReady
+	models := map[string]config.ModelConfig{
+		model: {ConcurrencyLimit: limit},
+	}
+	s := NewFIFO("test", logmon.NewWriter(io.Discard), &stubPlanner{}, config.FifoConfig{}, models, eff)
+	return s, eff
+}
+
+// TestFIFO_ConcurrencyLimit_RejectsOverLimit verifies that a request arriving
+// while the model is at capacity gets an error grant instead of being served,
+// and that a new request succeeds once an in-flight one completes.
+func TestFIFO_ConcurrencyLimit_RejectsOverLimit(t *testing.T) {
+	s, eff := newFIFOWithLimit(t, "a", 1)
+
+	// First request: served (inFlight 0 → 1).
+	s.OnRequest(req("a"))
+	if got := eff.served("a"); got != 1 {
+		t.Fatalf("served(a)=%d want 1", got)
+	}
+
+	// Second request while slot is occupied: rejected with HTTPError 429.
+	s.OnRequest(req("a"))
+	if got := eff.errored("a"); got != 1 {
+		t.Fatalf("errored(a)=%d want 1 (over-limit)", got)
+	}
+	var httpErr shared.HTTPError
+	if !errors.As(eff.grants[len(eff.grants)-1].err, &httpErr) {
+		t.Fatalf("err=%v want HTTPError", eff.grants[len(eff.grants)-1].err)
+	}
+	if httpErr.StatusCode() != http.StatusTooManyRequests {
+		t.Fatalf("StatusCode()=%d want 429", httpErr.StatusCode())
+	}
+	if httpErr.Header().Get("Retry-After") == "" {
+		t.Fatal("missing Retry-After header")
+	}
+
+	// After the in-flight request finishes, a new request succeeds.
+	s.OnServeDone(ServeDoneEvent{ModelID: "a"})
+	s.OnRequest(req("a"))
+	if got := eff.served("a"); got != 2 {
+		t.Fatalf("served(a)=%d want 2 after drain", got)
+	}
+}
+
+// TestFIFO_ConcurrencyLimit_DefaultIsTen verifies that a model without an
+// explicit ConcurrencyLimit gets the default cap of 10.
+func TestFIFO_ConcurrencyLimit_DefaultIsTen(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateReady
+	// nil models → every model gets defaultConcurrencyLimit (10).
+	s := newFIFO(&stubPlanner{}, eff)
+
+	for i := 0; i < 10; i++ {
+		s.OnRequest(req("a"))
+	}
+	if got := eff.served("a"); got != 10 {
+		t.Fatalf("served(a)=%d want 10 (default limit)", got)
+	}
+
+	// 11th request is rejected.
+	s.OnRequest(req("a"))
+	if got := eff.errored("a"); got != 1 {
+		t.Fatalf("errored(a)=%d want 1 (over default limit)", got)
+	}
+}
+
+// TestFIFO_ConcurrencyLimit_CustomLimit verifies a ConcurrencyLimit greater
+// than zero overrides the default.
+func TestFIFO_ConcurrencyLimit_CustomLimit(t *testing.T) {
+	s, eff := newFIFOWithLimit(t, "a", 2)
+
+	s.OnRequest(req("a"))
+	s.OnRequest(req("a"))
+	s.OnRequest(req("a"))
+
+	if got := eff.served("a"); got != 2 {
+		t.Fatalf("served(a)=%d want 2 (custom limit)", got)
+	}
+	if got := eff.errored("a"); got != 1 {
+		t.Fatalf("errored(a)=%d want 1 (over custom limit)", got)
+	}
+}
+
+// TestFIFO_ConcurrencyLimit_SwapWaiters verifies that when more swap waiters
+// exist than the concurrency limit, excess waiters are rejected on swap
+// completion rather than exceeding the limit.
+func TestFIFO_ConcurrencyLimit_SwapWaiters(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	models := map[string]config.ModelConfig{
+		"a": {ConcurrencyLimit: 2},
+	}
+	s := NewFIFO("test", logmon.NewWriter(io.Discard), &stubPlanner{}, config.FifoConfig{}, models, eff)
+
+	// Three requests arrive while model is loading: one starts swap, two join.
+	s.OnRequest(req("a"))
+	s.OnRequest(req("a"))
+	s.OnRequest(req("a"))
+
+	if got := eff.startsFor("a"); got != 1 {
+		t.Fatalf("StartSwap(a)=%d want 1", got)
+	}
+
+	// Swap completes: two served (limit), one rejected.
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+
+	if got := eff.served("a"); got != 2 {
+		t.Fatalf("served(a)=%d want 2 (limit on swap completion)", got)
+	}
+	if got := eff.errored("a"); got != 1 {
+		t.Fatalf("errored(a)=%d want 1 (excess waiter rejected)", got)
+	}
+}

internal/router/scheduler/scheduler.go

@@ -0,0 +1,143 @@
+// Package scheduler contains the request-scheduling strategies used by the
+// router's baseRouter. A Scheduler owns the queue, in-flight tracking, and the
+// decision tree for when to start a swap versus queue a request. The baseRouter
+// owns the channels, run loop, and process machinery, and exposes the
+// side-effects a scheduler needs through the Effects interface.
+//
+// Splitting these apart lets the scheduling strategy be swapped out
+// independently of both the process machinery (baseRouter) and the eviction
+// policy (Swapper). FIFO is the first and currently only implementation.
+package scheduler
+
+import (
+	"context"
+	"fmt"
+	"net/http"
+	"time"
+
+	"github.com/mostlygeek/llama-swap/internal/config"
+	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"github.com/mostlygeek/llama-swap/internal/process"
+	"github.com/mostlygeek/llama-swap/internal/shared"
+)
+
+// ErrModelNotFound is granted to callers whose model is not handled by this
+// router. It is an alias for shared.ErrNoLocalModelFound.
+var ErrModelNotFound = shared.ErrNoLocalModelFound
+
+// Swapper is the eviction policy: it decides which running models must be
+// stopped before a target can serve. It is orthogonal to the scheduling
+// strategy — any Scheduler works with any Swapper.
+type Swapper interface {
+	// EvictionFor returns running model IDs that must be stopped before
+	// target can serve. running is the complete set the scheduler considers
+	// live: every process that is not stopped, unioned with the targets of
+	// in-flight swaps the scheduler has already committed to (which are not yet
+	// visible in process state). The planner does not inspect process state
+	// itself. Pure decision; must not log.
+	EvictionFor(target string, running []string) []string
+
+	// OnSwapStart runs once at the start of every swap, with the same running
+	// set EvictionFor was given for this decision. Planners may log their
+	// decision here at whatever verbosity they choose.
+	OnSwapStart(target string, running []string)
+}
+
+// Scheduler decides what happens to each event the router's run loop receives.
+// All methods run on that single run-loop goroutine, so implementations need no
+// internal locking for their own state.
+type Scheduler interface {
+	// OnRequest handles one incoming ServeHTTP request.
+	OnRequest(req HandlerReq)
+	// OnCancel handles a request whose client has disconnected before it was
+	// granted. The scheduler must remove the request from its queue and from
+	// any in-flight swap's waiters so it never triggers a model load or grant
+	// for a caller that is no longer there.
+	OnCancel(req HandlerReq)
+	// OnSwapDone handles a swap goroutine reporting completion.
+	OnSwapDone(ev SwapDone)
+	// OnServeDone handles a tracked ServeHTTP finishing (in-flight decrement).
+	OnServeDone(ev ServeDoneEvent)
+	// OnUnload reconciles scheduler state for an unload, stops the targeted
+	// processes via Effects, and drains the queue. It must block until the
+	// targeted processes have stopped.
+	OnUnload(targets []string, timeout time.Duration)
+	// OnShutdown grants err to every waiter the scheduler still holds (active
+	// swap waiters and queued requests). Process teardown is the baseRouter's
+	// responsibility.
+	OnShutdown(err error)
+}
+
+// Effects is implemented by the baseRouter. The scheduler calls back through it
+// for every side-effect: inspecting process state, launching swaps, responding
+// to callers, and stopping processes.
+type Effects interface {
+	// ModelState returns the current state of a model's process. ok is false
+	// when the model is not handled by this router.
+	ModelState(modelID string) (process.ProcessState, bool)
+	// RunningModels returns the state of every process that is not stopped or
+	// shut down, keyed by model ID. The scheduler uses it to build the running
+	// set it hands the Swapper.
+	RunningModels() map[string]process.ProcessState
+	// StartSwap launches the swap goroutine for modelID, stopping evict first.
+	StartSwap(modelID string, evict []string)
+	// GrantError responds to a caller with an error.
+	GrantError(req HandlerReq, err error)
+	// GrantServe hands a caller the wrapped handler for modelID and reports
+	// whether the caller was still there to receive it. The scheduler bumps
+	// its in-flight count only when this returns true.
+	GrantServe(req HandlerReq, modelID string) bool
+	// StopProcesses stops the named processes in parallel and blocks until all
+	// have stopped. Unknown IDs are skipped.
+	StopProcesses(timeout time.Duration, ids []string)
+}
+
+// New returns a Scheduler selected by conf.Routing.Scheduler.Use, configured from
+// conf and bound to the given planner and effects. Supported values are "fifo"
+// (throughput-oriented, batches same-model requests) and "serial" (strict
+// one-model-at-a-time, exact arrival order).
+//
+// The deployment default is applied by config loading (LoadConfig sets Use to
+// "serial" when unset). The "" fallback here is the library default and remains
+// "fifo" so callers that build a Config directly keep the original behavior.
+func New(conf config.Config, name string, logger *logmon.Monitor, planner Swapper, eff Effects) (Scheduler, error) {
+	use := conf.Routing.Scheduler.Use
+	if use == "" {
+		use = "fifo"
+	}
+	switch use {
+	case "fifo":
+		return NewFIFO(name, logger, planner, conf.Routing.Scheduler.Settings.Fifo, conf.Models, eff), nil
+	case "serial":
+		// Serial ignores the group planner: it always evicts every other model.
+		return NewSerial(name, logger, eff), nil
+	default:
+		return nil, fmt.Errorf("unsupported scheduler type: %q", use)
+	}
+}
+
+// HandlerReq is one in-flight ServeHTTP request waiting for a routing decision.
+type HandlerReq struct {
+	Model      string
+	Ctx        context.Context
+	Respond    chan HandlerResp
+	PositionCh chan int
+}
+
+// HandlerResp is the routing decision returned to a HandlerReq's caller: either
+// a handler to serve with, or an error.
+type HandlerResp struct {
+	HandleFunc http.HandlerFunc
+	Err        error
+}
+
+// SwapDone is reported by a swap goroutine when its target is ready (or failed).
+type SwapDone struct {
+	ModelID string
+	Err     error
+}
+
+// ServeDoneEvent is reported when a tracked ServeHTTP handler returns.
+type ServeDoneEvent struct {
+	ModelID string
+}

internal/router/scheduler/serial.go

@@ -0,0 +1,253 @@
+package scheduler
+
+import (
+	"fmt"
+	"sort"
+	"time"
+
+	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"github.com/mostlygeek/llama-swap/internal/process"
+)
+
+// Serial is a strict one-model-at-a-time scheduler. Unlike FIFO it never reorders
+// or batches: requests run in exact arrival order and at most one request runs at
+// any instant. When the next request targets a model other than the one loaded,
+// every other running model is evicted and the target is loaded before it runs,
+// so a single model occupies memory at a time — at the cost of throughput.
+//
+// Example: A B C A is served as A B C A. The final A reloads its model even
+// though it ran first, because B and C displaced it in between. (FIFO, by
+// contrast, would batch the two A requests: A A B C.)
+//
+// Serial ignores group/eviction policy entirely: it always evicts every other
+// running model, regardless of how groups are configured. That is what makes the
+// single-model guarantee a property of the scheduler rather than of the config.
+//
+// Like FIFO, every method runs on the router's single run-loop goroutine, so no
+// internal locking is needed.
+type Serial struct {
+	name    string
+	logger  *logmon.Monitor
+	effects Effects
+
+	// queued holds requests in strict arrival order. It is never reordered.
+	queued []HandlerReq
+
+	// active is the one request currently being processed (loading or serving),
+	// or nil when idle. phase is meaningful only while active != nil.
+	active *HandlerReq
+	phase  serialPhase
+}
+
+// serialPhase is the lifecycle stage of the active request.
+type serialPhase int
+
+const (
+	phaseIdle     serialPhase = iota
+	phaseSwapping             // waiting for OnSwapDone for active.Model
+	phaseServing              // waiting for OnServeDone for active.Model
+)
+
+// NewSerial builds a Serial scheduler. It takes no Swapper: eviction is always
+// "stop every other running model", so the group planner is not consulted.
+func NewSerial(name string, logger *logmon.Monitor, eff Effects) *Serial {
+	return &Serial{
+		name:    name,
+		logger:  logger,
+		effects: eff,
+	}
+}
+
+// OnRequest validates the model and appends the request to the tail of the queue,
+// then tries to start the next job. Unknown models fail immediately.
+func (s *Serial) OnRequest(req HandlerReq) {
+	if _, ok := s.effects.ModelState(req.Model); !ok {
+		s.logger.Debugf("%s: model %s not handled by this router", s.name, req.Model)
+		s.effects.GrantError(req, ErrModelNotFound)
+		return
+	}
+	s.queued = append(s.queued, req)
+	broadcastQueuePositions(s.queued)
+	s.startNext()
+}
+
+// startNext begins processing the head of the queue when nothing is active. It
+// fast-paths a request whose model is already the sole loaded-and-ready process;
+// otherwise it launches a swap that evicts every other running model first. The
+// loop skips over requests for models that vanished (e.g. a config reload) and
+// requests whose caller disconnected before they could be served.
+func (s *Serial) startNext() {
+	if s.active != nil {
+		return // a job is already loading or serving
+	}
+	for len(s.queued) > 0 {
+		req := s.queued[0]
+		s.queued = s.queued[1:]
+		broadcastQueuePositions(s.queued)
+
+		state, ok := s.effects.ModelState(req.Model)
+		if !ok {
+			s.effects.GrantError(req, ErrModelNotFound)
+			continue
+		}
+
+		r := req
+		s.active = &r
+
+		evict := s.otherRunning(req.Model)
+		if state == process.StateReady && len(evict) == 0 {
+			// Already loaded and the only model running — serve immediately.
+			s.logger.Debugf("%s: serving model %s (already loaded)", s.name, req.Model)
+			if s.serve() {
+				return
+			}
+			continue // caller gone; pick the next request
+		}
+
+		s.logger.Debugf("%s: swapping to model %s, evicting %v", s.name, req.Model, evict)
+		s.phase = phaseSwapping
+		s.effects.StartSwap(req.Model, evict)
+		return
+	}
+}
+
+// serve hands the active request its tracked handler. It returns true when the
+// request is now serving (await OnServeDone); false when the caller had already
+// disconnected, in which case active is cleared so the next job can start.
+func (s *Serial) serve() bool {
+	if s.effects.GrantServe(*s.active, s.active.Model) {
+		s.phase = phaseServing
+		return true
+	}
+	s.logger.Debugf("%s: caller for model %s gone before serve", s.name, s.active.Model)
+	s.active = nil
+	s.phase = phaseIdle
+	return false
+}
+
+// OnSwapDone fires when the load for the active request completes. On success the
+// request is served; on failure its caller receives the error and the queue
+// advances. A SwapDone that does not match the active load (e.g. its request was
+// unloaded or cancelled mid-load) is ignored.
+func (s *Serial) OnSwapDone(ev SwapDone) {
+	if s.active == nil || s.phase != phaseSwapping || s.active.Model != ev.ModelID {
+		return
+	}
+	if ev.Err != nil {
+		s.logger.Debugf("%s: swap for model %s failed: %v", s.name, ev.ModelID, ev.Err)
+		s.effects.GrantError(*s.active, ev.Err)
+		s.active = nil
+		s.phase = phaseIdle
+		s.startNext()
+		return
+	}
+	if !s.serve() {
+		s.startNext() // caller vanished while the model loaded; move on
+	}
+}
+
+// OnServeDone fires when the active request's handler returns. The slot is freed
+// and the next queued request begins.
+func (s *Serial) OnServeDone(ev ServeDoneEvent) {
+	if s.active == nil || s.phase != phaseServing {
+		return
+	}
+	s.active = nil
+	s.phase = phaseIdle
+	s.startNext()
+}
+
+// OnCancel removes a disconnected client's request from the queue. A request that
+// is already active is left to finish: if it was loading, OnSwapDone's serve()
+// will find the caller gone (GrantServe false) and advance; if it was serving,
+// its handler returns normally and reaches OnServeDone.
+func (s *Serial) OnCancel(req HandlerReq) {
+	if len(s.queued) == 0 {
+		return
+	}
+	kept := s.queued[:0]
+	removed := false
+	for _, q := range s.queued {
+		if q.Respond == req.Respond {
+			removed = true
+			continue
+		}
+		kept = append(kept, q)
+	}
+	s.queued = kept
+	if removed {
+		s.logger.Debugf("%s: cancelled request for model %s pruned from queue", s.name, req.Model)
+		broadcastQueuePositions(s.queued)
+	}
+}
+
+// OnUnload reconciles state for an unload, stops the targeted processes, and
+// advances the queue. It mirrors the FIFO contract: queued requests for unloaded
+// models are failed; an active *loading* request for an unloaded model is failed
+// (its swap goroutine is left to finish and its SwapDone is then ignored); an
+// active *serving* request is left for its handler to end when StopProcesses
+// kills the upstream. The Stop is synchronous so callers of Unload can rely on
+// the processes being stopped on return.
+func (s *Serial) OnUnload(targets []string, timeout time.Duration) {
+	unloadErr := fmt.Errorf("%s: model unloaded", s.name)
+
+	targetSet := make(map[string]bool, len(targets))
+	for _, id := range targets {
+		targetSet[id] = true
+	}
+
+	if s.active != nil && s.phase == phaseSwapping && targetSet[s.active.Model] {
+		s.effects.GrantError(*s.active, unloadErr)
+		s.active = nil
+		s.phase = phaseIdle
+	}
+
+	if len(s.queued) > 0 {
+		kept := s.queued[:0]
+		for _, q := range s.queued {
+			if targetSet[q.Model] {
+				s.effects.GrantError(q, unloadErr)
+				continue
+			}
+			kept = append(kept, q)
+		}
+		s.queued = kept
+		broadcastQueuePositions(s.queued)
+	}
+
+	s.effects.StopProcesses(timeout, targets)
+
+	// A still-serving active request advances via OnServeDone when its killed
+	// handler returns; only start the next job when nothing is active now.
+	if s.active == nil {
+		s.startNext()
+	}
+}
+
+// OnShutdown grants err to every request the scheduler still holds: an active
+// loading request and all queued requests. A serving request is torn down with
+// its process by the baseRouter.
+func (s *Serial) OnShutdown(err error) {
+	if s.active != nil && s.phase == phaseSwapping {
+		s.effects.GrantError(*s.active, err)
+		s.active = nil
+		s.phase = phaseIdle
+	}
+	for _, q := range s.queued {
+		s.effects.GrantError(q, err)
+	}
+	s.queued = nil
+}
+
+// otherRunning returns every running model except target, sorted for
+// deterministic eviction.
+func (s *Serial) otherRunning(target string) []string {
+	var out []string
+	for id := range s.effects.RunningModels() {
+		if id != target {
+			out = append(out, id)
+		}
+	}
+	sort.Strings(out)
+	return out
+}

internal/router/scheduler/serial_test.go

@@ -0,0 +1,391 @@
+package scheduler
+
+import (
+	"errors"
+	"io"
+	"testing"
+	"time"
+
+	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"github.com/mostlygeek/llama-swap/internal/process"
+)
+
+// Serial methods all run on the router's single run-loop goroutine, so these
+// tests drive them directly and synchronously, reusing fakeEffects and the
+// req/reqCh helpers from fifo_test.go. A load completes via OnSwapDone and a
+// served request finishes via OnServeDone — the events the run loop delivers.
+
+func newSerial(eff Effects) *Serial {
+	return NewSerial("test", logmon.NewWriter(io.Discard), eff)
+}
+
+// lastStart returns the most recent StartSwap record.
+func lastStart(t *testing.T, eff *fakeEffects) startRec {
+	t.Helper()
+	if len(eff.starts) == 0 {
+		t.Fatal("no StartSwap recorded")
+	}
+	return eff.starts[len(eff.starts)-1]
+}
+
+func sameSet(a, b []string) bool {
+	if len(a) != len(b) {
+		return false
+	}
+	m := map[string]int{}
+	for _, x := range a {
+		m[x]++
+	}
+	for _, x := range b {
+		m[x]--
+	}
+	for _, v := range m {
+		if v != 0 {
+			return false
+		}
+	}
+	return true
+}
+
+// servedOrder returns the model IDs of every successful serve grant in order.
+func servedOrder(eff *fakeEffects) []string {
+	var out []string
+	for _, g := range eff.grants {
+		if g.err == nil && g.serve {
+			out = append(out, g.model)
+		}
+	}
+	return out
+}
+
+func TestSerial_FastPath_AlreadyLoaded(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateReady
+	s := newSerial(eff)
+
+	s.OnRequest(req("a"))
+
+	if got := len(eff.starts); got != 0 {
+		t.Errorf("StartSwap calls=%d want 0 (already loaded, no swap)", got)
+	}
+	if got := eff.served("a"); got != 1 {
+		t.Errorf("served(a)=%d want 1", got)
+	}
+}
+
+func TestSerial_ColdStart_LoadsThenServes(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	s := newSerial(eff)
+
+	s.OnRequest(req("a"))
+	if got := eff.startsFor("a"); got != 1 {
+		t.Fatalf("StartSwap(a)=%d want 1", got)
+	}
+	if got := eff.served("a"); got != 0 {
+		t.Errorf("served(a)=%d want 0 before load completes", got)
+	}
+
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+	if got := eff.served("a"); got != 1 {
+		t.Errorf("served(a)=%d want 1 after load", got)
+	}
+}
+
+func TestSerial_UnknownModel(t *testing.T) {
+	eff := newFakeEffects() // no states => unknown
+	s := newSerial(eff)
+
+	s.OnRequest(req("ghost"))
+
+	if len(eff.starts) != 0 {
+		t.Errorf("StartSwap calls=%d want 0", len(eff.starts))
+	}
+	if eff.errored("ghost") != 1 {
+		t.Fatalf("errored(ghost)=%d want 1", eff.errored("ghost"))
+	}
+	if !errors.Is(eff.grants[0].err, ErrModelNotFound) {
+		t.Errorf("err=%v want ErrModelNotFound", eff.grants[0].err)
+	}
+}
+
+func TestSerial_EvictsEveryOtherModel(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["x"] = process.StateReady // already running
+	eff.states["y"] = process.StateReady // also running (e.g. left over)
+	eff.states["a"] = process.StateStopped
+	s := newSerial(eff)
+
+	s.OnRequest(req("a"))
+
+	st := lastStart(t, eff)
+	if st.model != "a" {
+		t.Fatalf("loading %s want a", st.model)
+	}
+	if !sameSet(st.evict, []string{"x", "y"}) {
+		t.Errorf("evict=%v want [x y] (serial evicts ALL other models)", st.evict)
+	}
+}
+
+// TestSerial_OneJobAtATime verifies a second request waits while the first is
+// serving, and only starts after the first finishes.
+func TestSerial_OneJobAtATime(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateReady
+	eff.states["b"] = process.StateStopped
+	s := newSerial(eff)
+
+	s.OnRequest(req("a")) // served immediately
+	s.OnRequest(req("b")) // must wait — a is serving
+
+	if got := eff.startsFor("b"); got != 0 {
+		t.Fatalf("StartSwap(b)=%d want 0 while a is serving", got)
+	}
+	if got := eff.served("a"); got != 1 {
+		t.Fatalf("served(a)=%d want 1", got)
+	}
+
+	// a finishes -> b may now load (evicting a).
+	s.OnServeDone(ServeDoneEvent{ModelID: "a"})
+	if got := eff.startsFor("b"); got != 1 {
+		t.Fatalf("StartSwap(b)=%d want 1 after a finished", got)
+	}
+	if st := lastStart(t, eff); !sameSet(st.evict, []string{"a"}) {
+		t.Errorf("b evict=%v want [a]", st.evict)
+	}
+}
+
+// TestSerial_SameModelConsecutive_NoReload verifies back-to-back requests for the
+// already-loaded model run without a reload, one after another.
+func TestSerial_SameModelConsecutive_NoReload(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	s := newSerial(eff)
+
+	s.OnRequest(req("a")) // cold load
+	s.OnRequest(req("a")) // queued behind the first
+
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"}) // first serves
+	if got := eff.served("a"); got != 1 {
+		t.Fatalf("served(a)=%d want 1 (one at a time)", got)
+	}
+
+	s.OnServeDone(ServeDoneEvent{ModelID: "a"}) // first done -> second serves
+	if got := eff.served("a"); got != 2 {
+		t.Fatalf("served(a)=%d want 2", got)
+	}
+	if got := eff.startsFor("a"); got != 1 {
+		t.Errorf("StartSwap(a)=%d want 1 (second request must not reload)", got)
+	}
+}
+
+// TestSerial_StrictArrivalOrder is the core guarantee: qwen36, qwen35, sdxl,
+// qwen36 execute in EXACTLY that order with evictions between each model switch,
+// including reloading qwen36 at the end even though it ran first.
+func TestSerial_StrictArrivalOrder(t *testing.T) {
+	eff := newFakeEffects()
+	for _, m := range []string{"qwen36", "qwen35", "sdxl"} {
+		eff.states[m] = process.StateStopped
+	}
+	s := newSerial(eff)
+
+	for _, m := range []string{"qwen36", "qwen35", "sdxl", "qwen36"} {
+		s.OnRequest(req(m))
+	}
+
+	// Only the first job starts loading; the rest wait their turn.
+	if len(eff.starts) != 1 || eff.starts[0].model != "qwen36" {
+		t.Fatalf("starts=%+v want only [qwen36] loading first", eff.starts)
+	}
+
+	// step completes the current model's load+serve and returns control to the
+	// scheduler, which must start the next queued model.
+	step := func(model string, wantEvict []string) {
+		t.Helper()
+		st := lastStart(t, eff)
+		if st.model != model {
+			t.Fatalf("loading %q want %q", st.model, model)
+		}
+		if !sameSet(st.evict, wantEvict) {
+			t.Fatalf("loading %q evict=%v want %v", model, st.evict, wantEvict)
+		}
+		// Simulate the eviction + load actually happening.
+		for _, e := range st.evict {
+			eff.states[e] = process.StateStopped
+		}
+		eff.states[model] = process.StateReady
+		s.OnSwapDone(SwapDone{ModelID: model})
+		s.OnServeDone(ServeDoneEvent{ModelID: model})
+	}
+
+	step("qwen36", nil)                // cold load, nothing else running
+	step("qwen35", []string{"qwen36"}) // evict qwen36
+	step("sdxl", []string{"qwen35"})   // evict qwen35
+	step("qwen36", []string{"sdxl"})   // RELOAD qwen36, evict sdxl
+
+	want := []string{"qwen36", "qwen35", "sdxl", "qwen36"}
+	if got := servedOrder(eff); !sameOrder(got, want) {
+		t.Fatalf("serve order=%v want %v", got, want)
+	}
+}
+
+func sameOrder(a, b []string) bool {
+	if len(a) != len(b) {
+		return false
+	}
+	for i := range a {
+		if a[i] != b[i] {
+			return false
+		}
+	}
+	return true
+}
+
+func TestSerial_SwapError_FailsCallerAndAdvances(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	s := newSerial(eff)
+
+	s.OnRequest(req("a"))
+	s.OnRequest(req("b")) // queued behind a
+
+	// a's load fails: its caller is errored and b proceeds.
+	s.OnSwapDone(SwapDone{ModelID: "a", Err: errors.New("boom")})
+	if eff.errored("a") != 1 {
+		t.Fatalf("errored(a)=%d want 1", eff.errored("a"))
+	}
+	if got := eff.startsFor("b"); got != 1 {
+		t.Fatalf("StartSwap(b)=%d want 1 after a's load failed", got)
+	}
+}
+
+// TestSerial_GrantServeFalse_Advances verifies that when the active request's
+// caller has disconnected by serve time, the queue advances to the next request.
+func TestSerial_GrantServeFalse_Advances(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	eff.serveResult["a"] = false // a's caller is gone by grant time
+	s := newSerial(eff)
+
+	s.OnRequest(req("a"))
+	s.OnRequest(req("b")) // queued
+
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"}) // grant fails -> advance to b
+
+	if got := eff.served("a"); got != 0 {
+		t.Errorf("served(a)=%d want 0 (caller gone)", got)
+	}
+	if got := eff.startsFor("b"); got != 1 {
+		t.Fatalf("StartSwap(b)=%d want 1 (advanced after gone caller)", got)
+	}
+}
+
+func TestSerial_OnCancel_QueuedRequest(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	s := newSerial(eff)
+
+	s.OnRequest(reqCh("a")) // starts loading a
+	cancelled := reqCh("b")
+	s.OnRequest(cancelled) // queued behind a
+	if len(s.queued) != 1 {
+		t.Fatalf("queued=%d want 1", len(s.queued))
+	}
+
+	s.OnCancel(cancelled)
+	if len(s.queued) != 0 {
+		t.Fatalf("queued=%d want 0 after cancel", len(s.queued))
+	}
+
+	// a completes; b is gone, so nothing starts for it.
+	eff.states["a"] = process.StateReady
+	s.OnSwapDone(SwapDone{ModelID: "a"})
+	s.OnServeDone(ServeDoneEvent{ModelID: "a"})
+	if got := eff.startsFor("b"); got != 0 {
+		t.Errorf("StartSwap(b)=%d want 0 (cancelled before its turn)", got)
+	}
+}
+
+func TestSerial_OnShutdown_FailsQueuedAndActiveLoad(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	eff.states["c"] = process.StateStopped
+	s := newSerial(eff)
+
+	s.OnRequest(req("a")) // active (loading)
+	s.OnRequest(req("b")) // queued
+	s.OnRequest(req("c")) // queued
+
+	s.OnShutdown(errors.New("shutting down"))
+
+	if got := eff.errored(""); got != 3 {
+		t.Errorf("error grants=%d want 3 (active load + 2 queued)", got)
+	}
+	if len(s.queued) != 0 {
+		t.Errorf("queued=%d want 0 after shutdown", len(s.queued))
+	}
+}
+
+// TestSerial_OnUnload_WhileServing verifies that unloading the model that is
+// actively serving does not strand the queue: OnUnload stops the process but
+// leaves the active request to end via OnServeDone, which then advances.
+func TestSerial_OnUnload_WhileServing(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateReady
+	eff.states["b"] = process.StateStopped
+	s := newSerial(eff)
+
+	s.OnRequest(req("a")) // served immediately (a ready)
+	s.OnRequest(req("b")) // queued behind a
+	if got := eff.served("a"); got != 1 {
+		t.Fatalf("served(a)=%d want 1", got)
+	}
+
+	// Unload a while it is serving: the process is stopped, but the queue must
+	// not advance yet — the active serve is still outstanding.
+	s.OnUnload([]string{"a"}, time.Second)
+	if len(eff.stops) != 1 || !sameSet(eff.stops[0].ids, []string{"a"}) {
+		t.Errorf("StopProcesses=%+v want one call stopping [a]", eff.stops)
+	}
+	if got := eff.startsFor("b"); got != 0 {
+		t.Fatalf("StartSwap(b)=%d want 0 before the serving request ends", got)
+	}
+
+	// The killed handler returns -> OnServeDone advances to b.
+	eff.states["a"] = process.StateStopped
+	s.OnServeDone(ServeDoneEvent{ModelID: "a"})
+	if got := eff.startsFor("b"); got != 1 {
+		t.Fatalf("StartSwap(b)=%d want 1 after the serving request ended", got)
+	}
+}
+
+func TestSerial_OnUnload_DropsQueuedAndStops(t *testing.T) {
+	eff := newFakeEffects()
+	eff.states["a"] = process.StateStopped
+	eff.states["b"] = process.StateStopped
+	s := newSerial(eff)
+
+	s.OnRequest(req("a")) // active (loading a)
+	s.OnRequest(req("b")) // queued
+
+	// Unload a: its active load is failed and a is stopped.
+	s.OnUnload([]string{"a"}, time.Second)
+
+	if eff.errored("a") != 1 {
+		t.Errorf("errored(a)=%d want 1 (active load failed)", eff.errored("a"))
+	}
+	if len(eff.stops) != 1 || !sameSet(eff.stops[0].ids, []string{"a"}) {
+		t.Errorf("StopProcesses=%+v want one call stopping [a]", eff.stops)
+	}
+	// b was queued and not unloaded; with a's load cancelled it now starts.
+	if got := eff.startsFor("b"); got != 1 {
+		t.Errorf("StartSwap(b)=%d want 1 after unload advanced the queue", got)
+	}
+}

internal/server/api.go

@@ -2,6 +2,7 @@ package server
 
 import (
 	"encoding/json"
+	"fmt"
 	"net/http"
 	"sort"
 	"strings"
@@ -9,19 +10,127 @@ import (
 
 	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/event"
-	"github.com/mostlygeek/llama-swap/internal/router"
+	"github.com/mostlygeek/llama-swap/internal/process"
 	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
+// apiUnloadTimeout is used by the API endpoints to stop processes
+const apiUnloadTimeout = 10 * time.Second
+
 // modelRecord is one entry in the OpenAI-compatible /v1/models listing.
 type modelRecord struct {
-	ID          string         `json:"id"`
-	Object      string         `json:"object"`
-	Created     int64          `json:"created"`
-	OwnedBy     string         `json:"owned_by"`
-	Name        string         `json:"name,omitempty"`
-	Description string         `json:"description,omitempty"`
-	Meta        map[string]any `json:"meta,omitempty"`
+	ID                  string         `json:"id"`
+	Object              string         `json:"object"`
+	Created             int64          `json:"created"`
+	OwnedBy             string         `json:"owned_by"`
+	Name                string         `json:"name,omitempty"`
+	Description         string         `json:"description,omitempty"`
+	Architecture        map[string]any `json:"architecture,omitempty"`
+	Capabilities        map[string]any `json:"capabilities,omitempty"`
+	SupportedParameters []string       `json:"supported_parameters,omitempty"`
+	ContextLength       int            `json:"context_length,omitempty"`
+	Meta                map[string]any `json:"meta,omitempty"`
+}
+
+// cappedMetadataKeys are top-level /v1/models fields produced by the
+// capabilities renderer. If a model's metadata block defines any of these
+// keys, the renderer's values win and the metadata keys are dropped.
+var cappedMetadataKeys = map[string]struct{}{
+	"architecture":         {},
+	"capabilities":         {},
+	"supported_parameters": {},
+	"context_length":       {},
+}
+
+// renderCapabilities converts a model's capabilities config into additional
+// /v1/models fields. Returns zero values when caps.Empty() is true.
+func renderCapabilities(caps config.ModelCapConfig) (arch map[string]any, capsMap map[string]any, params []string, ctxLen int) {
+	if caps.Empty() {
+		return
+	}
+
+	hasIn := len(caps.In) > 0
+	hasOut := len(caps.Out) > 0
+
+	if hasIn || hasOut {
+		arch = make(map[string]any)
+	}
+	if hasIn {
+		arch["input_modalities"] = caps.In
+	}
+	if hasOut {
+		arch["output_modalities"] = caps.Out
+	}
+	if hasIn && hasOut {
+		arch["modality"] = strings.Join(caps.In, "+") + "->" + strings.Join(caps.Out, "+")
+	}
+
+	// Build capabilities map only if there's something to put in it.
+	if hasIn || hasOut || caps.Tools || caps.Reranker {
+		capsMap = make(map[string]any)
+	}
+
+	if hasIn {
+		if contains(caps.In, "image") {
+			capsMap["vision"] = true
+		}
+	}
+	if hasIn && hasOut {
+		if contains(caps.In, "audio") && contains(caps.Out, "text") {
+			capsMap["audio_transcriptions"] = true
+		}
+		if contains(caps.In, "text") && contains(caps.Out, "audio") {
+			capsMap["audio_speech"] = true
+		}
+		if contains(caps.In, "text") && contains(caps.Out, "image") {
+			capsMap["image_generation"] = true
+		}
+		if contains(caps.In, "image") && contains(caps.Out, "image") {
+			capsMap["image_to_image"] = true
+		}
+	}
+
+	if caps.Tools {
+		capsMap["function_calling"] = true
+		params = []string{"tools", "tool_choice"}
+	}
+
+	if caps.Reranker {
+		capsMap["reranker"] = true
+	}
+
+	if caps.Context > 0 {
+		ctxLen = caps.Context
+	}
+
+	return
+}
+
+// contains reports whether s is present in ss.
+func contains(ss []string, s string) bool {
+	for _, v := range ss {
+		if v == s {
+			return true
+		}
+	}
+	return false
+}
+
+// filterCappedMetadata returns metadata with renderer-owned keys removed.
+func filterCappedMetadata(md map[string]any) map[string]any {
+	if len(md) == 0 {
+		return nil
+	}
+	filtered := make(map[string]any, len(md))
+	for k, v := range md {
+		if _, capped := cappedMetadataKeys[k]; !capped {
+			filtered[k] = v
+		}
+	}
+	if len(filtered) == 0 {
+		return nil
+	}
+	return filtered
 }
 
 // handleListModels serves the OpenAI-compatible model listing: local models
@@ -30,7 +139,7 @@ func (s *Server) handleListModels(w http.ResponseWriter, r *http.Request) {
 	created := time.Now().Unix()
 	data := make([]modelRecord, 0, len(s.cfg.Models))
 
-	newRecord := func(id, name, description string, metadata map[string]any) modelRecord {
+	newRecord := func(id, name, description string, metadata map[string]any, caps config.ModelCapConfig) modelRecord {
 		rec := modelRecord{
 			ID:          id,
 			Object:      "model",
@@ -39,6 +148,10 @@ func (s *Server) handleListModels(w http.ResponseWriter, r *http.Request) {
 			Name:        strings.TrimSpace(name),
 			Description: strings.TrimSpace(description),
 		}
+		rec.Architecture, rec.Capabilities, rec.SupportedParameters, rec.ContextLength = renderCapabilities(caps)
+		if !caps.Empty() {
+			metadata = filterCappedMetadata(metadata)
+		}
 		if len(metadata) > 0 {
 			rec.Meta = map[string]any{"llamaswap": metadata}
 		}
@@ -49,12 +162,12 @@ func (s *Server) handleListModels(w http.ResponseWriter, r *http.Request) {
 		if mc.Unlisted {
 			continue
 		}
-		data = append(data, newRecord(id, mc.Name, mc.Description, mc.Metadata))
+		data = append(data, newRecord(id, mc.Name, mc.Description, mc.Metadata, mc.Capabilities))
 
 		if s.cfg.IncludeAliasesInList {
 			for _, alias := range mc.Aliases {
 				if alias := strings.TrimSpace(alias); alias != "" {
-					data = append(data, newRecord(alias, mc.Name, mc.Description, mc.Metadata))
+					data = append(data, newRecord(alias, mc.Name, mc.Description, mc.Metadata, mc.Capabilities))
 				}
 			}
 		}
@@ -62,7 +175,7 @@ func (s *Server) handleListModels(w http.ResponseWriter, r *http.Request) {
 
 	for peerID, peer := range s.cfg.Peers {
 		for _, modelID := range peer.Models {
-			data = append(data, newRecord(modelID, peerID+": "+modelID, "", map[string]any{"peerID": peerID}))
+			data = append(data, newRecord(modelID, peerID+": "+modelID, "", map[string]any{"peerID": peerID}, config.ModelCapConfig{}))
 		}
 	}
 
@@ -94,7 +207,7 @@ type runningModel struct {
 // handleUnload stops every running local process. Peer models are remote and
 // unaffected.
 func (s *Server) handleUnload(w http.ResponseWriter, r *http.Request) {
-	s.local.Unload(0)
+	s.local.Unload(apiUnloadTimeout)
 	w.WriteHeader(http.StatusOK)
 	w.Write([]byte("OK"))
 }
@@ -160,7 +273,7 @@ func (s *Server) startPreload() {
 			if err != nil {
 				continue
 			}
-			req = req.WithContext(router.SetContext(req.Context(), router.ReqContextData{Model: modelID, ModelID: modelID}))
+			req = req.WithContext(shared.SetContext(req.Context(), shared.ReqContextData{Model: modelID, ModelID: modelID, Metadata: make(map[string]string)}))
 
 			dw := &discardResponseWriter{status: http.StatusOK}
 			s.local.ServeHTTP(dw, req)
@@ -203,9 +316,9 @@ func handleUpstreamRedirect(w http.ResponseWriter, r *http.Request) {
 func (s *Server) handleUpstream(w http.ResponseWriter, r *http.Request) {
 	upstreamPath := r.PathValue("upstreamPath")
 
-	searchName, modelID, remainingPath, found := findModelInPath(s.cfg, "/"+upstreamPath)
+	searchName, modelID, remainingPath, found := shared.FindModelInPath(s.cfg, "/"+upstreamPath)
 	if !found {
-		router.SendResponse(w, r, http.StatusNotFound, "model not found")
+		shared.SendResponse(w, r, http.StatusNotFound, "model not found")
 		return
 	}
 
@@ -227,7 +340,29 @@ func (s *Server) handleUpstream(w http.ResponseWriter, r *http.Request) {
 	// Strip the /upstream/<model> prefix before forwarding.
 	r.URL.Path = remainingPath
 	// Pin the resolved model so the router skips body/query extraction.
-	*r = *r.WithContext(router.SetContext(r.Context(), router.ReqContextData{Model: searchName, ModelID: modelID}))
+	*r = *r.WithContext(shared.SetContext(r.Context(), shared.ReqContextData{Model: searchName, ModelID: modelID, Metadata: make(map[string]string)}))
+
+	// If the path matches an upstream.ignorePaths entry and the model is
+	// not already loaded, refuse the request without triggering a swap. The
+	// server was not able to process the response because the model was not
+	// already loaded.
+	for _, re := range s.cfg.Upstream.IgnorePaths {
+		if !re.MatchString(remainingPath) {
+			continue
+		}
+		if s.local.Handles(modelID) {
+			state, ok := s.local.RunningModels()[modelID]
+			if !ok || state != process.StateReady {
+				shared.SendResponse(w, r, http.StatusConflict,
+					fmt.Sprintf("model %s is not loaded; path matches upstream.ignorePaths", modelID))
+				return
+			}
+		}
+		// Either the model is already loaded (no swap would be triggered)
+		// or this is a peer model (peer proxying never swaps). Fall through
+		// to normal dispatch.
+		break
+	}
 
 	switch {
 	case s.local.Handles(modelID):
@@ -235,32 +370,6 @@ func (s *Server) handleUpstream(w http.ResponseWriter, r *http.Request) {
 	case s.peer.Handles(modelID):
 		s.peer.ServeHTTP(w, r)
 	default:
-		router.SendResponse(w, r, http.StatusNotFound, "no router for model "+modelID)
+		shared.SendResponse(w, r, http.StatusNotFound, "no router for model "+modelID)
 	}
 }
-
-// findModelInPath walks a slash-separated path, building up segments until one
-// matches a configured model. This resolves model names that contain slashes
-// (e.g. "author/model"). Returns the matched name, its real model ID, the
-// remaining path, and whether a match was found.
-func findModelInPath(cfg config.Config, path string) (searchName, realName, remainingPath string, found bool) {
-	parts := strings.Split(strings.TrimSpace(path), "/")
-	name := ""
-
-	for i, part := range parts {
-		if part == "" {
-			continue
-		}
-		if name == "" {
-			name = part
-		} else {
-			name = name + "/" + part
-		}
-
-		if modelID, ok := cfg.RealModelName(name); ok {
-			return name, modelID, "/" + strings.Join(parts[i+1:], "/"), true
-		}
-	}
-
-	return "", "", "", false
-}

internal/server/api_test.go

@@ -2,11 +2,17 @@ package server
 
 import (
 	"encoding/json"
+	"io"
 	"net/http"
 	"net/http/httptest"
+	"regexp"
+	"strings"
 	"testing"
 
 	"github.com/mostlygeek/llama-swap/internal/config"
+	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"github.com/mostlygeek/llama-swap/internal/process"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
 func TestServer_HandleListModels(t *testing.T) {
@@ -78,6 +84,7 @@ func TestServer_HandleListModels_Aliases(t *testing.T) {
 
 func TestServer_FindModelInPath(t *testing.T) {
 	cfg := config.Config{Models: map[string]config.ModelConfig{
+		"author":       {},
 		"author/model": {},
 		"simple":       {},
 	}}
@@ -91,13 +98,14 @@ func TestServer_FindModelInPath(t *testing.T) {
 		{"/simple/v1/chat", "simple", "/v1/chat", true},
 		{"/author/model/v1/chat", "author/model", "/v1/chat", true},
 		{"/author/model", "author/model", "/", true},
+		{"/author/v1/chat", "author", "/v1/chat", true},
 		{"/missing/v1", "", "", false},
 		{"/", "", "", false},
 	}
 	for _, c := range cases {
-		name, _, rem, found := findModelInPath(cfg, c.path)
+		name, _, rem, found := shared.FindModelInPath(cfg, c.path)
 		if found != c.wantFound || name != c.wantName || (found && rem != c.wantRem) {
-			t.Errorf("findModelInPath(%q) = (%q,%q,%v), want (%q,%q,%v)",
+			t.Errorf("FindModelInPath(%q) = (%q,%q,%v), want (%q,%q,%v)",
 				c.path, name, rem, found, c.wantName, c.wantRem, c.wantFound)
 		}
 	}
@@ -133,6 +141,165 @@ func TestServer_HandleUpstream(t *testing.T) {
 	})
 }
 
+func upstreamMetricsServer(response string) *Server {
+	cfg := config.Config{Models: map[string]config.ModelConfig{"m1": {}}}
+	proxylog := logmon.NewWriter(io.Discard)
+	s := &Server{
+		cfg:         cfg,
+		muxlog:      logmon.NewWriter(io.Discard),
+		proxylog:    proxylog,
+		upstreamlog: logmon.NewWriter(io.Discard),
+		inflight:    &inflightCounter{},
+		metrics:     newMetricsMonitor(proxylog, 10, 0),
+		local:       newStubRouter([]string{"m1"}, response),
+		peer:        newStubRouter(nil, ""),
+	}
+	s.routes()
+	return s
+}
+
+func TestServer_HandleUpstream_IgnorePaths(t *testing.T) {
+	// Compile a pattern that matches static asset suffixes.
+	pattern := regexp.MustCompile(`.*\.(js|json|css|png|gif|jpg|jpeg|txt)$`)
+
+	t.Run("matched path, model not loaded, returns 409", func(t *testing.T) {
+		local := newStubRouter([]string{"m1"}, "upstream-body")
+		// running is nil/empty: model is not in RunningModels() => not loaded.
+		s := newTestServer(local, newStubRouter(nil, ""))
+		s.cfg = config.Config{
+			Models: map[string]config.ModelConfig{"m1": {}},
+			Upstream: config.UpstreamConfig{
+				IgnorePaths: []*regexp.Regexp{pattern},
+			},
+		}
+
+		w := httptest.NewRecorder()
+		s.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/upstream/m1/foo.js", nil))
+
+		if w.Code != http.StatusConflict {
+			t.Fatalf("status = %d, want %d (body=%q)", w.Code, http.StatusConflict, w.Body.String())
+		}
+		if !strings.Contains(w.Body.String(), "not loaded") {
+			t.Errorf("body = %q, want it to contain 'not loaded'", w.Body.String())
+		}
+	})
+
+	t.Run("matched path, model already loaded, serves normally", func(t *testing.T) {
+		local := newStubRouter([]string{"m1"}, "upstream-body")
+		local.running = map[string]process.ProcessState{"m1": process.StateReady}
+		s := newTestServer(local, newStubRouter(nil, ""))
+		s.cfg = config.Config{
+			Models: map[string]config.ModelConfig{"m1": {}},
+			Upstream: config.UpstreamConfig{
+				IgnorePaths: []*regexp.Regexp{pattern},
+			},
+		}
+
+		w := httptest.NewRecorder()
+		s.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/upstream/m1/foo.js", nil))
+
+		if w.Code != http.StatusOK || w.Body.String() != "upstream-body" {
+			t.Fatalf("status=%d body=%q, want 200 'upstream-body'", w.Code, w.Body.String())
+		}
+	})
+
+	t.Run("non-matched path, model not loaded, serves normally", func(t *testing.T) {
+		local := newStubRouter([]string{"m1"}, "upstream-body")
+		s := newTestServer(local, newStubRouter(nil, ""))
+		s.cfg = config.Config{
+			Models: map[string]config.ModelConfig{"m1": {}},
+			Upstream: config.UpstreamConfig{
+				IgnorePaths: []*regexp.Regexp{pattern},
+			},
+		}
+
+		w := httptest.NewRecorder()
+		s.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/upstream/m1/v1/chat/completions", nil))
+
+		if w.Code != http.StatusOK || w.Body.String() != "upstream-body" {
+			t.Fatalf("status=%d body=%q, want 200 'upstream-body'", w.Code, w.Body.String())
+		}
+	})
+
+	t.Run("matched path, peer model, serves normally", func(t *testing.T) {
+		// Peer routers do not appear via RunningModels on the local router;
+		// they should fall through to normal dispatch without 409.
+		local := newStubRouter(nil, "")
+		peer := newStubRouter([]string{"m1"}, "peer-body")
+		s := newTestServer(local, peer)
+		s.cfg = config.Config{
+			Models: map[string]config.ModelConfig{"m1": {}},
+			Upstream: config.UpstreamConfig{
+				IgnorePaths: []*regexp.Regexp{pattern},
+			},
+		}
+
+		w := httptest.NewRecorder()
+		s.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/upstream/m1/foo.js", nil))
+
+		if w.Code != http.StatusOK || w.Body.String() != "peer-body" {
+			t.Fatalf("status=%d body=%q, want 200 'peer-body'", w.Code, w.Body.String())
+		}
+	})
+}
+
+func TestServer_HandleUpstream_MetricsRecordsSupportedPath(t *testing.T) {
+	resp := `{"usage":{"prompt_tokens":3,"completion_tokens":5}}`
+	s := upstreamMetricsServer(resp)
+
+	w := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodPost, "/upstream/m1/v1/chat/completions", strings.NewReader(`{}`))
+	req.Header.Set("Content-Type", "application/json")
+	s.ServeHTTP(w, req)
+
+	if w.Code != http.StatusOK || w.Body.String() != resp {
+		t.Fatalf("status=%d body=%q", w.Code, w.Body.String())
+	}
+	entries := s.metrics.getMetrics()
+	if len(entries) != 1 {
+		t.Fatalf("want 1 metrics entry, got %d", len(entries))
+	}
+	if entries[0].Model != "m1" {
+		t.Errorf("model = %q, want m1", entries[0].Model)
+	}
+	if entries[0].ReqPath != "/v1/chat/completions" {
+		t.Errorf("req_path = %q, want /v1/chat/completions", entries[0].ReqPath)
+	}
+	if entries[0].Tokens.InputTokens != 3 || entries[0].Tokens.OutputTokens != 5 {
+		t.Errorf("tokens = %+v, want input=3 output=5", entries[0].Tokens)
+	}
+}
+
+func TestServer_HandleUpstream_MetricsSkipsUnsupportedPath(t *testing.T) {
+	s := upstreamMetricsServer("ok")
+
+	w := httptest.NewRecorder()
+	req := httptest.NewRequest(http.MethodPost, "/upstream/m1/probe", strings.NewReader(`{}`))
+	req.Header.Set("Content-Type", "application/json")
+	s.ServeHTTP(w, req)
+
+	if w.Code != http.StatusOK || w.Body.String() != "ok" {
+		t.Fatalf("status=%d body=%q", w.Code, w.Body.String())
+	}
+	if len(s.metrics.getMetrics()) != 0 {
+		t.Errorf("want no metrics entries for unsupported path, got %d", len(s.metrics.getMetrics()))
+	}
+}
+
+func TestServer_HandleUpstream_MetricsSkipsGET(t *testing.T) {
+	s := upstreamMetricsServer(`{"usage":{}}`)
+
+	w := httptest.NewRecorder()
+	s.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/upstream/m1/v1/chat/completions", nil))
+
+	if w.Code != http.StatusOK {
+		t.Fatalf("status=%d", w.Code)
+	}
+	if len(s.metrics.getMetrics()) != 0 {
+		t.Errorf("want no metrics entries for GET upstream, got %d", len(s.metrics.getMetrics()))
+	}
+}
+
 func TestServer_HandleMetrics_Unavailable(t *testing.T) {
 	s := newTestServer(newStubRouter(nil, ""), newStubRouter(nil, ""))
 
@@ -157,3 +324,262 @@ func TestServer_Redirects(t *testing.T) {
 		}
 	}
 }
+
+func TestServer_HandleListModels_Capabilities(t *testing.T) {
+	newServer := func(mc config.ModelConfig) *Server {
+		s := newTestServer(newStubRouter(nil, ""), newStubRouter(nil, ""))
+		s.cfg = config.Config{Models: map[string]config.ModelConfig{"m": mc}}
+		return s
+	}
+	getModel := func(t *testing.T, s *Server) modelRecord {
+		t.Helper()
+		w := httptest.NewRecorder()
+		s.ServeHTTP(w, httptest.NewRequest(http.MethodGet, "/v1/models", nil))
+		var resp struct {
+			Data []modelRecord `json:"data"`
+		}
+		if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+			t.Fatalf("decode: %v", err)
+		}
+		if len(resp.Data) != 1 {
+			t.Fatalf("expected 1 model, got %d", len(resp.Data))
+		}
+		return resp.Data[0]
+	}
+
+	t.Run("all_fields", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{
+				In:      []string{"text", "image"},
+				Out:     []string{"text", "audio"},
+				Tools:   true,
+				Context: 100000,
+			},
+		}))
+		if m.Architecture == nil {
+			t.Fatal("architecture is nil")
+		}
+		if !anySliceStrEqual(m.Architecture["input_modalities"], []string{"text", "image"}) {
+			t.Errorf("input_modalities = %v", m.Architecture["input_modalities"])
+		}
+		if !anySliceStrEqual(m.Architecture["output_modalities"], []string{"text", "audio"}) {
+			t.Errorf("output_modalities = %v", m.Architecture["output_modalities"])
+		}
+		if m.Architecture["modality"] != "text+image->text+audio" {
+			t.Errorf("modality = %v", m.Architecture["modality"])
+		}
+		if m.Capabilities == nil || m.Capabilities["vision"] != true {
+			t.Errorf("vision = %v", m.Capabilities)
+		}
+		if m.Capabilities["audio_speech"] != true {
+			t.Errorf("audio_speech = %v", m.Capabilities["audio_speech"])
+		}
+		if m.Capabilities["function_calling"] != true {
+			t.Errorf("function_calling = %v", m.Capabilities["function_calling"])
+		}
+		if !stringSliceEqual(m.SupportedParameters, []string{"tools", "tool_choice"}) {
+			t.Errorf("supported_parameters = %v", m.SupportedParameters)
+		}
+		if m.ContextLength != 100000 {
+			t.Errorf("context_length = %d", m.ContextLength)
+		}
+	})
+
+	t.Run("in_only", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{In: []string{"text", "image"}},
+		}))
+		if m.Architecture == nil {
+			t.Fatal("architecture is nil")
+		}
+		if _, ok := m.Architecture["output_modalities"]; ok {
+			t.Error("should not have output_modalities")
+		}
+		if _, ok := m.Architecture["modality"]; ok {
+			t.Error("should not have modality")
+		}
+		if m.Capabilities == nil || m.Capabilities["vision"] != true {
+			t.Error("expected vision: true")
+		}
+		if m.SupportedParameters != nil {
+			t.Error("should not have supported_parameters")
+		}
+		if m.ContextLength != 0 {
+			t.Error("should not have context_length")
+		}
+	})
+
+	t.Run("out_only", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{Out: []string{"audio"}},
+		}))
+		if m.Architecture == nil {
+			t.Fatal("architecture is nil")
+		}
+		if _, ok := m.Architecture["input_modalities"]; ok {
+			t.Error("should not have input_modalities")
+		}
+		if len(m.Capabilities) > 0 {
+			t.Errorf("expected no capabilities, got %v", m.Capabilities)
+		}
+	})
+
+	t.Run("tools", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{Tools: true},
+		}))
+		if m.Capabilities == nil || m.Capabilities["function_calling"] != true {
+			t.Error("expected function_calling: true")
+		}
+		if !stringSliceEqual(m.SupportedParameters, []string{"tools", "tool_choice"}) {
+			t.Errorf("supported_parameters = %v", m.SupportedParameters)
+		}
+		if m.Architecture != nil {
+			t.Error("should not have architecture")
+		}
+	})
+
+	t.Run("reranker", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{Reranker: true},
+		}))
+		if m.Capabilities == nil || m.Capabilities["reranker"] != true {
+			t.Error("expected reranker: true")
+		}
+		if m.Architecture != nil {
+			t.Error("should not have architecture")
+		}
+	})
+
+	t.Run("context", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{Context: 32768},
+		}))
+		if m.ContextLength != 32768 {
+			t.Errorf("context_length = %d", m.ContextLength)
+		}
+		if m.Architecture != nil {
+			t.Error("should not have architecture")
+		}
+	})
+
+	t.Run("audio_transcriptions", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{In: []string{"audio"}, Out: []string{"text"}},
+		}))
+		if m.Capabilities == nil || m.Capabilities["audio_transcriptions"] != true {
+			t.Error("expected audio_transcriptions: true")
+		}
+	})
+
+	t.Run("image_generation", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{In: []string{"text"}, Out: []string{"image"}},
+		}))
+		if m.Capabilities == nil || m.Capabilities["image_generation"] != true {
+			t.Error("expected image_generation: true")
+		}
+	})
+
+	t.Run("image_to_image", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{In: []string{"image"}, Out: []string{"image"}},
+		}))
+		if m.Capabilities == nil || m.Capabilities["image_to_image"] != true {
+			t.Error("expected image_to_image: true")
+		}
+	})
+
+	t.Run("empty_skip", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{}))
+		if m.Architecture != nil {
+			t.Error("should not have architecture")
+		}
+		if m.Capabilities != nil {
+			t.Error("should not have capabilities")
+		}
+		if m.SupportedParameters != nil {
+			t.Error("should not have supported_parameters")
+		}
+		if m.ContextLength != 0 {
+			t.Error("should not have context_length")
+		}
+	})
+
+	t.Run("metadata_precedence", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Capabilities: config.ModelCapConfig{In: []string{"text"}},
+			Metadata: map[string]any{
+				"architecture":   "should-be-dropped",
+				"custom_field":   "should-remain",
+				"capabilities":   "also-dropped",
+				"other_metadata": "also-remain",
+			},
+		}))
+		if m.Architecture == nil || m.Architecture["input_modalities"] == nil {
+			t.Fatal("architecture should be rendered, not from metadata")
+		}
+		if m.Meta == nil || m.Meta["llamaswap"] == nil {
+			t.Fatal("meta.llamaswap should exist")
+		}
+		meta := m.Meta["llamaswap"].(map[string]any)
+		if _, ok := meta["architecture"]; ok {
+			t.Error("architecture should be filtered from metadata")
+		}
+		if _, ok := meta["custom_field"]; !ok {
+			t.Error("custom_field should remain in metadata")
+		}
+	})
+
+	t.Run("metadata_passthrough_no_caps", func(t *testing.T) {
+		m := getModel(t, newServer(config.ModelConfig{
+			Metadata: map[string]any{
+				"architecture":   "preserved",
+				"context_length": 4096,
+				"capabilities":   "preserved",
+				"custom_field":   "preserved",
+			},
+		}))
+		if m.Architecture != nil {
+			t.Error("should not have architecture when caps is empty")
+		}
+		if m.Meta == nil || m.Meta["llamaswap"] == nil {
+			t.Fatal("meta.llamaswap should exist")
+		}
+		meta := m.Meta["llamaswap"].(map[string]any)
+		if _, ok := meta["architecture"]; !ok {
+			t.Error("architecture should be preserved in metadata when caps is empty")
+		}
+		if _, ok := meta["context_length"]; !ok {
+			t.Error("context_length should be preserved in metadata when caps is empty")
+		}
+	})
+}
+
+func stringSliceEqual(a, b []string) bool {
+	if len(a) != len(b) {
+		return false
+	}
+	for i := range a {
+		if a[i] != b[i] {
+			return false
+		}
+	}
+	return true
+}
+
+func anySliceStrEqual(v any, want []string) bool {
+	arr, ok := v.([]any)
+	if !ok {
+		return false
+	}
+	if len(arr) != len(want) {
+		return false
+	}
+	for i := range arr {
+		if s, ok := arr[i].(string); !ok || s != want[i] {
+			return false
+		}
+	}
+	return true
+}

internal/server/apigroup.go

@@ -12,19 +12,19 @@ import (
 
 	"github.com/mostlygeek/llama-swap/internal/event"
 	"github.com/mostlygeek/llama-swap/internal/perf"
-	"github.com/mostlygeek/llama-swap/internal/router"
 	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
 // apiModel is one entry in the /api/events modelStatus payload.
 type apiModel struct {
-	Id          string   `json:"id"`
-	Name        string   `json:"name"`
-	Description string   `json:"description"`
-	State       string   `json:"state"`
-	Unlisted    bool     `json:"unlisted"`
-	PeerID      string   `json:"peerID"`
-	Aliases     []string `json:"aliases,omitempty"`
+	Id           string         `json:"id"`
+	Name         string         `json:"name"`
+	Description  string         `json:"description"`
+	State        string         `json:"state"`
+	Unlisted     bool           `json:"unlisted"`
+	PeerID       string         `json:"peerID"`
+	Aliases      []string       `json:"aliases,omitempty"`
+	Capabilities map[string]any `json:"capabilities,omitempty"`
 }
 
 // modelStatus returns every configured model joined with its current process
@@ -45,13 +45,15 @@ func (s *Server) modelStatus() []apiModel {
 		if st, ok := running[id]; ok {
 			state = string(st)
 		}
+		_, capsMap, _, _ := renderCapabilities(mc.Capabilities)
 		models = append(models, apiModel{
-			Id:          id,
-			Name:        mc.Name,
-			Description: mc.Description,
-			State:       state,
-			Unlisted:    mc.Unlisted,
-			Aliases:     mc.Aliases,
+			Id:           id,
+			Name:         mc.Name,
+			Description:  mc.Description,
+			State:        state,
+			Unlisted:     mc.Unlisted,
+			Aliases:      mc.Aliases,
+			Capabilities: capsMap,
 		})
 	}
 
@@ -66,7 +68,7 @@ func (s *Server) modelStatus() []apiModel {
 
 // handleAPIUnloadAll stops every running local process.
 func (s *Server) handleAPIUnloadAll(w http.ResponseWriter, r *http.Request) {
-	s.local.Unload(0)
+	s.local.Unload(apiUnloadTimeout)
 	w.Header().Set("Content-Type", "application/json")
 	json.NewEncoder(w).Encode(map[string]string{"msg": "ok"})
 }
@@ -76,14 +78,14 @@ func (s *Server) handleAPIUnloadModel(w http.ResponseWriter, r *http.Request) {
 	requested := strings.TrimPrefix(r.PathValue("model"), "/")
 	realName, found := s.cfg.RealModelName(requested)
 	if !found {
-		router.SendResponse(w, r, http.StatusNotFound, "model not found")
+		shared.SendResponse(w, r, http.StatusNotFound, "model not found")
 		return
 	}
 	if !s.local.Handles(realName) {
-		router.SendResponse(w, r, http.StatusNotFound, "no local server found for requested model")
+		shared.SendResponse(w, r, http.StatusNotFound, "no local server found for requested model")
 		return
 	}
-	s.local.Unload(0, realName)
+	s.local.Unload(apiUnloadTimeout, realName)
 	w.WriteHeader(http.StatusOK)
 	w.Write([]byte("OK"))
 }
@@ -92,7 +94,7 @@ func (s *Server) handleAPIUnloadModel(w http.ResponseWriter, r *http.Request) {
 func (s *Server) handleAPIMetrics(w http.ResponseWriter, r *http.Request) {
 	data, err := s.metrics.getMetricsJSON()
 	if err != nil {
-		router.SendResponse(w, r, http.StatusInternalServerError, "failed to get metrics")
+		shared.SendResponse(w, r, http.StatusInternalServerError, "failed to get metrics")
 		return
 	}
 	w.Header().Set("Content-Type", "application/json")
@@ -103,7 +105,9 @@ func (s *Server) handleAPIMetrics(w http.ResponseWriter, r *http.Request) {
 // filtered to samples after the ?after=<RFC3339> timestamp.
 func (s *Server) handleAPIPerformance(w http.ResponseWriter, r *http.Request) {
 	if s.perf == nil {
-		router.SendResponse(w, r, http.StatusServiceUnavailable, "performance monitor not available")
+		w.Header().Set("Content-Type", "application/json")
+		w.WriteHeader(http.StatusServiceUnavailable)
+		json.NewEncoder(w).Encode(map[string]bool{"enabled": false})
 		return
 	}
 
@@ -112,7 +116,7 @@ func (s *Server) handleAPIPerformance(w http.ResponseWriter, r *http.Request) {
 	if afterStr := r.URL.Query().Get("after"); afterStr != "" {
 		after, err := time.Parse(time.RFC3339, afterStr)
 		if err != nil {
-			router.SendResponse(w, r, http.StatusBadRequest, "invalid 'after' timestamp, use RFC3339 format")
+			shared.SendResponse(w, r, http.StatusBadRequest, "invalid 'after' timestamp, use RFC3339 format")
 			return
 		}
 		filteredSys := make([]perf.SysStat, 0, len(sysStats))
@@ -134,6 +138,7 @@ func (s *Server) handleAPIPerformance(w http.ResponseWriter, r *http.Request) {
 
 	w.Header().Set("Content-Type", "application/json")
 	json.NewEncoder(w).Encode(map[string]any{
+		"enabled":   true,
 		"sys_stats": sysStats,
 		"gpu_stats": gpuStats,
 	})
@@ -153,19 +158,19 @@ func (s *Server) handleAPIVersion(w http.ResponseWriter, r *http.Request) {
 func (s *Server) handleAPICapture(w http.ResponseWriter, r *http.Request) {
 	id, err := strconv.Atoi(r.PathValue("id"))
 	if err != nil {
-		router.SendResponse(w, r, http.StatusBadRequest, "invalid capture ID")
+		shared.SendResponse(w, r, http.StatusBadRequest, "invalid capture ID")
 		return
 	}
 
 	capture := s.metrics.getCaptureByID(id)
 	if capture == nil {
-		router.SendResponse(w, r, http.StatusNotFound, "capture not found")
+		shared.SendResponse(w, r, http.StatusNotFound, "capture not found")
 		return
 	}
 
 	jsonBytes, err := json.Marshal(capture)
 	if err != nil {
-		router.SendResponse(w, r, http.StatusInternalServerError, "failed to marshal capture")
+		shared.SendResponse(w, r, http.StatusInternalServerError, "failed to marshal capture")
 		return
 	}
 	w.Header().Set("Content-Type", "application/json")
@@ -198,7 +203,7 @@ func (s *Server) handleAPIEvents(w http.ResponseWriter, r *http.Request) {
 
 	flusher, ok := w.(http.Flusher)
 	if !ok {
-		router.SendResponse(w, r, http.StatusInternalServerError, "streaming unsupported")
+		shared.SendResponse(w, r, http.StatusInternalServerError, "streaming unsupported")
 		return
 	}
 

internal/server/auth.go

@@ -1,19 +1,17 @@
 package server
 
 import (
-	"encoding/base64"
 	"net/http"
 	"strings"
 
 	"github.com/mostlygeek/llama-swap/internal/chain"
 	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/mostlygeek/llama-swap/internal/router"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
 // CreateAuthMiddleware returns middleware that validates API keys when the
 // config declares any. It accepts the key via Authorization: Bearer,
-// Authorization: Basic (password field), or x-api-key. On success the auth
-// headers are stripped so they never leak to upstream. When no keys are
+// Authorization: Basic (password field), or x-api-key. When no keys are
 // configured the middleware is a pass-through.
 func CreateAuthMiddleware(cfg config.Config) chain.Middleware {
 	keys := cfg.RequiredAPIKeys
@@ -22,7 +20,7 @@ func CreateAuthMiddleware(cfg config.Config) chain.Middleware {
 			return next
 		}
 		return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-			provided := extractAPIKey(r)
+			provided := shared.ExtractAPIKey(r)
 
 			valid := false
 			for _, key := range keys {
@@ -33,41 +31,29 @@ func CreateAuthMiddleware(cfg config.Config) chain.Middleware {
 			}
 			if !valid {
 				w.Header().Set("WWW-Authenticate", `Basic realm="llama-swap"`)
-				router.SendResponse(w, r, http.StatusUnauthorized, "unauthorized: invalid or missing API key")
+				shared.SendResponse(w, r, http.StatusUnauthorized, "unauthorized: invalid or missing API key")
 				return
 			}
 
-			r.Header.Del("Authorization")
-			r.Header.Del("x-api-key")
 			next.ServeHTTP(w, r)
 		})
 	}
 }
 
-// extractAPIKey pulls a candidate API key from the request, preferring Basic,
-// then Bearer, then x-api-key.
-func extractAPIKey(r *http.Request) string {
-	var bearerKey, basicKey string
-	if auth := r.Header.Get("Authorization"); auth != "" {
-		if strings.HasPrefix(auth, "Bearer ") {
-			bearerKey = strings.TrimPrefix(auth, "Bearer ")
-		} else if strings.HasPrefix(auth, "Basic ") {
-			encoded := strings.TrimPrefix(auth, "Basic ")
-			if decoded, err := base64.StdEncoding.DecodeString(encoded); err == nil {
-				if parts := strings.SplitN(string(decoded), ":", 2); len(parts) == 2 {
-					basicKey = parts[1] // password field is the API key
-				}
+// CreateRequestContextMiddleware returns middleware that extracts model and
+// auth info from the request into the context. Requests where no model can be
+// identified are rejected with a 404.
+func CreateRequestContextMiddleware(cfg config.Config) chain.Middleware {
+	return func(next http.Handler) http.Handler {
+		return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+			data, err := shared.FetchContext(r, cfg)
+			if err != nil {
+				shared.SendError(w, r, shared.ErrNoModelInContext)
+				return
 			}
-		}
-	}
-
-	switch {
-	case basicKey != "":
-		return basicKey
-	case bearerKey != "":
-		return bearerKey
-	default:
-		return r.Header.Get("x-api-key")
+			_ = data
+			next.ServeHTTP(w, r)
+		})
 	}
 }
 

internal/server/auth_test.go

@@ -1,48 +1,14 @@
 package server
 
 import (
-	"encoding/base64"
 	"net/http"
 	"net/http/httptest"
+	"strings"
 	"testing"
 
 	"github.com/mostlygeek/llama-swap/internal/config"
 )
 
-func TestServer_ExtractAPIKey(t *testing.T) {
-	basicHeader := func(user, pass string) string {
-		return "Basic " + base64.StdEncoding.EncodeToString([]byte(user+":"+pass))
-	}
-	cases := []struct {
-		name string
-		auth string
-		xapi string
-		want string
-	}{
-		{"none", "", "", ""},
-		{"bearer", "Bearer tok123", "", "tok123"},
-		{"basic", basicHeader("user", "pw-key"), "", "pw-key"},
-		{"x-api-key", "", "xkey", "xkey"},
-		{"basic beats bearer", basicHeader("u", "bk"), "", "bk"},
-		{"bearer beats x-api-key", "Bearer btok", "xkey", "btok"},
-		{"malformed basic falls back to x-api-key", "Basic !!!notbase64", "xkey", "xkey"},
-	}
-	for _, c := range cases {
-		t.Run(c.name, func(t *testing.T) {
-			r := httptest.NewRequest(http.MethodGet, "/", nil)
-			if c.auth != "" {
-				r.Header.Set("Authorization", c.auth)
-			}
-			if c.xapi != "" {
-				r.Header.Set("x-api-key", c.xapi)
-			}
-			if got := extractAPIKey(r); got != c.want {
-				t.Errorf("extractAPIKey() = %q, want %q", got, c.want)
-			}
-		})
-	}
-}
-
 func TestServer_SanitizeAccessControlRequestHeaders(t *testing.T) {
 	cases := []struct {
 		in   string
@@ -74,11 +40,42 @@ func TestServer_IsTokenChar(t *testing.T) {
 	}
 }
 
+func TestServer_RequestContextMiddleware(t *testing.T) {
+	cfg := config.Config{
+		Models: map[string]config.ModelConfig{
+			"llama3": {},
+		},
+	}
+
+	final := http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.WriteHeader(http.StatusOK)
+	})
+
+	mw := CreateRequestContextMiddleware(cfg)
+
+	t.Run("known model passes through", func(t *testing.T) {
+		r := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(`{"model":"llama3"}`))
+		r.Header.Set("Content-Type", "application/json")
+		w := httptest.NewRecorder()
+		mw(final).ServeHTTP(w, r)
+		if w.Code != http.StatusOK {
+			t.Errorf("status = %d, want 200", w.Code)
+		}
+	})
+
+	t.Run("missing model returns 404", func(t *testing.T) {
+		r := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(`{}`))
+		r.Header.Set("Content-Type", "application/json")
+		w := httptest.NewRecorder()
+		mw(final).ServeHTTP(w, r)
+		if w.Code != http.StatusNotFound {
+			t.Errorf("status = %d, want 404", w.Code)
+		}
+	})
+}
+
 func TestServer_AuthMiddleware(t *testing.T) {
 	final := http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		if r.Header.Get("Authorization") != "" || r.Header.Get("x-api-key") != "" {
-			t.Error("auth headers leaked to upstream")
-		}
 		w.WriteHeader(http.StatusOK)
 	})
 

internal/server/concurrency.go

@@ -1,55 +0,0 @@
-package server
-
-import (
-	"net/http"
-
-	"golang.org/x/sync/semaphore"
-
-	"github.com/mostlygeek/llama-swap/internal/chain"
-	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/mostlygeek/llama-swap/internal/router"
-)
-
-// defaultConcurrencyLimit caps simultaneous in-flight requests per model when
-// the model config leaves concurrencyLimit unset. Matches the legacy
-// proxy.Process default.
-const defaultConcurrencyLimit = 10
-
-// CreateConcurrencyMiddleware returns middleware that limits simultaneous
-// model-dispatched requests per model. Each model gets a semaphore sized to
-// its concurrencyLimit (or defaultConcurrencyLimit). A request that cannot
-// immediately acquire a slot is rejected with 429. Models without a local
-// config entry (e.g. peer-routed models) are not limited.
-func CreateConcurrencyMiddleware(cfg config.Config) chain.Middleware {
-	semaphores := make(map[string]*semaphore.Weighted, len(cfg.Models))
-	for id, mc := range cfg.Models {
-		limit := defaultConcurrencyLimit
-		if mc.ConcurrencyLimit > 0 {
-			limit = mc.ConcurrencyLimit
-		}
-		semaphores[id] = semaphore.NewWeighted(int64(limit))
-	}
-
-	return func(next http.Handler) http.Handler {
-		return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-			data, err := router.FetchContext(r, cfg)
-			if err != nil {
-				router.SendError(w, r, router.ErrNoModelInContext)
-				return
-			}
-
-			// fall through for peer models
-			sem, ok := semaphores[data.ModelID]
-			if !ok {
-				next.ServeHTTP(w, r)
-				return
-			}
-			if !sem.TryAcquire(1) {
-				http.Error(w, "Too many requests", http.StatusTooManyRequests)
-				return
-			}
-			defer sem.Release(1)
-			next.ServeHTTP(w, r)
-		})
-	}
-}

internal/server/concurrency_test.go

@@ -1,75 +0,0 @@
-package server
-
-import (
-	"net/http"
-	"net/http/httptest"
-	"sync"
-	"testing"
-
-	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/mostlygeek/llama-swap/internal/router"
-)
-
-func concurrencyTestReq(model string) *http.Request {
-	r := httptest.NewRequest("GET", "/v1/chat/completions", nil)
-	return r.WithContext(router.SetContext(r.Context(), router.ReqContextData{Model: model, ModelID: model}))
-}
-
-func TestServer_ConcurrencyMiddleware_RejectsOverLimit(t *testing.T) {
-	cfg := config.Config{
-		Models: map[string]config.ModelConfig{
-			"m1": {ConcurrencyLimit: 1},
-		},
-	}
-
-	entered := make(chan struct{})
-	release := make(chan struct{})
-	var once sync.Once
-	final := http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		once.Do(func() { close(entered) })
-		<-release
-		w.WriteHeader(http.StatusOK)
-	})
-	h := CreateConcurrencyMiddleware(cfg)(final)
-
-	// First request occupies the only slot.
-	done := make(chan struct{})
-	go func() {
-		defer close(done)
-		h.ServeHTTP(httptest.NewRecorder(), concurrencyTestReq("m1"))
-	}()
-	<-entered
-
-	// Second concurrent request is rejected with 429.
-	w := httptest.NewRecorder()
-	h.ServeHTTP(w, concurrencyTestReq("m1"))
-	if w.Code != http.StatusTooManyRequests {
-		t.Fatalf("over-limit status = %d, want 429", w.Code)
-	}
-
-	// Once the slot frees, a new request succeeds.
-	close(release)
-	<-done
-	w = httptest.NewRecorder()
-	h.ServeHTTP(w, concurrencyTestReq("m1"))
-	if w.Code != http.StatusOK {
-		t.Fatalf("post-release status = %d, want 200", w.Code)
-	}
-}
-
-func TestServer_ConcurrencyMiddleware_UnconfiguredModelPassesThrough(t *testing.T) {
-	cfg := config.Config{Models: map[string]config.ModelConfig{}}
-
-	called := 0
-	final := http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		called++
-		w.WriteHeader(http.StatusOK)
-	})
-	h := CreateConcurrencyMiddleware(cfg)(final)
-
-	w := httptest.NewRecorder()
-	h.ServeHTTP(w, concurrencyTestReq("peer-model"))
-	if w.Code != http.StatusOK || called != 1 {
-		t.Fatalf("unconfigured model: status=%d called=%d, want 200/1", w.Code, called)
-	}
-}

internal/server/extras_test.go

@@ -1,10 +1,12 @@
 package server
 
 import (
+	"bufio"
 	"bytes"
 	"compress/flate"
 	"compress/gzip"
 	"io"
+	"net"
 	"net/http"
 	"net/http/httptest"
 	"testing"
@@ -75,6 +77,55 @@ func TestServer_BodyCopier_Flush(t *testing.T) {
 	}
 }
 
+// hijackRecorder is an httptest.ResponseRecorder that also implements
+// http.Hijacker, returning a pipe so Hijack forwarding can be exercised.
+type hijackRecorder struct {
+	*httptest.ResponseRecorder
+	conn net.Conn
+}
+
+func (h *hijackRecorder) Hijack() (net.Conn, *bufio.ReadWriter, error) {
+	return h.conn, bufio.NewReadWriter(bufio.NewReader(h.conn), bufio.NewWriter(h.conn)), nil
+}
+
+func TestServer_BodyCopier_Hijack(t *testing.T) {
+	t.Run("forwards to underlying hijacker", func(t *testing.T) {
+		client, server := net.Pipe()
+		defer client.Close()
+		defer server.Close()
+
+		bc := newBodyCopier(&hijackRecorder{httptest.NewRecorder(), server})
+		conn, _, err := bc.Hijack()
+		if err != nil {
+			t.Fatalf("Hijack: %v", err)
+		}
+		if conn != server {
+			t.Errorf("Hijack returned unexpected conn")
+		}
+	})
+
+	t.Run("errors when underlying writer is not a hijacker", func(t *testing.T) {
+		bc := newBodyCopier(httptest.NewRecorder())
+		if _, _, err := bc.Hijack(); err == nil {
+			t.Error("expected error hijacking a non-Hijacker ResponseWriter")
+		}
+	})
+}
+
+func TestServer_BodyCopier_SkipsBufferingOnUpgrade(t *testing.T) {
+	rec := httptest.NewRecorder()
+	bc := newBodyCopier(rec)
+	bc.WriteHeader(http.StatusSwitchingProtocols)
+	bc.Write([]byte("websocket frame bytes"))
+
+	if bc.body.Len() != 0 {
+		t.Errorf("upgrade body buffered = %q, want empty", bc.body.Bytes())
+	}
+	if got := rec.Body.String(); got != "websocket frame bytes" {
+		t.Errorf("client body = %q, want %q", got, "websocket frame bytes")
+	}
+}
+
 func TestServer_HeaderMapAndRedact(t *testing.T) {
 	h := http.Header{
 		"Content-Type":  {"application/json"},

internal/server/filters.go

@@ -11,7 +11,7 @@ import (
 
 	"github.com/mostlygeek/llama-swap/internal/chain"
 	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/mostlygeek/llama-swap/internal/router"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 	"github.com/tidwall/sjson"
 )
 
@@ -34,9 +34,9 @@ func CreateFilterMiddleware(cfg config.Config) chain.Middleware {
 				return
 			}
 
-			data, err := router.FetchContext(r, cfg)
+			data, err := shared.FetchContext(r, cfg)
 			if err != nil {
-				router.SendError(w, r, router.ErrNoModelInContext)
+				shared.SendError(w, r, shared.ErrNoModelInContext)
 				return
 			}
 
@@ -48,13 +48,13 @@ func CreateFilterMiddleware(cfg config.Config) chain.Middleware {
 
 			body, err := io.ReadAll(r.Body)
 			if err != nil {
-				router.SendResponse(w, r, http.StatusBadRequest, "could not read request body")
+				shared.SendResponse(w, r, http.StatusBadRequest, "could not read request body")
 				return
 			}
 
 			body, err = applyFilters(body, data.Model, useModelName, filters)
 			if err != nil {
-				router.SendResponse(w, r, http.StatusInternalServerError, err.Error())
+				shared.SendResponse(w, r, http.StatusInternalServerError, err.Error())
 				return
 			}
 
@@ -84,9 +84,9 @@ func CreateFormFilterMiddleware(cfg config.Config) chain.Middleware {
 				return
 			}
 
-			data, err := router.FetchContext(r, cfg)
+			data, err := shared.FetchContext(r, cfg)
 			if err != nil {
-				router.SendError(w, r, router.ErrNoModelInContext)
+				shared.SendError(w, r, shared.ErrNoModelInContext)
 				return
 			}
 
@@ -97,13 +97,13 @@ func CreateFormFilterMiddleware(cfg config.Config) chain.Middleware {
 			}
 
 			if err := r.ParseMultipartForm(32 << 20); err != nil {
-				router.SendResponse(w, r, http.StatusBadRequest, fmt.Sprintf("error parsing multipart form: %s", err.Error()))
+				shared.SendResponse(w, r, http.StatusBadRequest, fmt.Sprintf("error parsing multipart form: %s", err.Error()))
 				return
 			}
 
 			body, contentType, err := rewriteMultipartModel(r.MultipartForm, useModelName)
 			if err != nil {
-				router.SendResponse(w, r, http.StatusInternalServerError, err.Error())
+				shared.SendResponse(w, r, http.StatusInternalServerError, err.Error())
 				return
 			}
 

internal/server/log.go

@@ -1,6 +1,7 @@
 package server
 
 import (
+	"bufio"
 	"context"
 	"fmt"
 	"io"
@@ -13,7 +14,7 @@ import (
 	"github.com/mostlygeek/llama-swap/internal/chain"
 	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
-	"github.com/mostlygeek/llama-swap/internal/router"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
 // NewLoggers builds the proxy, upstream, and combined (mux) log monitors,
@@ -75,7 +76,7 @@ func (s *Server) getLogger(logMonitorID string) (*logmon.Monitor, error) {
 	case "upstream":
 		return s.upstreamlog, nil
 	default:
-		if _, modelID, _, found := findModelInPath(s.cfg, "/"+logMonitorID); found {
+		if _, modelID, _, found := shared.FindModelInPath(s.cfg, "/"+logMonitorID); found {
 			if log, ok := s.local.ProcessLogger(modelID); ok {
 				return log, nil
 			}
@@ -101,13 +102,13 @@ func (s *Server) handleLogStream(w http.ResponseWriter, r *http.Request) {
 
 	logger, err := s.getLogger(logMonitorID)
 	if err != nil {
-		router.SendResponse(w, r, http.StatusBadRequest, err.Error())
+		shared.SendResponse(w, r, http.StatusBadRequest, err.Error())
 		return
 	}
 
 	flusher, ok := w.(http.Flusher)
 	if !ok {
-		router.SendResponse(w, r, http.StatusInternalServerError, "streaming unsupported")
+		shared.SendResponse(w, r, http.StatusInternalServerError, "streaming unsupported")
 		return
 	}
 
@@ -150,7 +151,8 @@ var requestLogPathSkips = []string{"/wol-health", "/api/performance", "/metrics"
 
 // statusRecorder wraps an http.ResponseWriter to capture the response status
 // code and the number of body bytes written, so the access log can report
-// them. Flush is forwarded so streaming handlers (SSE) still work.
+// them. Flush is forwarded so streaming handlers (SSE) still work, and Hijack
+// is forwarded so httputil.ReverseProxy can upgrade websocket connections.
 type statusRecorder struct {
 	http.ResponseWriter
 	status int
@@ -174,6 +176,15 @@ func (sr *statusRecorder) Flush() {
 	}
 }
 
+// Hijack forwards to the underlying ResponseWriter so httputil.ReverseProxy can
+// take over the connection for websocket upgrades.
+func (sr *statusRecorder) Hijack() (net.Conn, *bufio.ReadWriter, error) {
+	if hj, ok := sr.ResponseWriter.(http.Hijacker); ok {
+		return hj.Hijack()
+	}
+	return nil, nil, fmt.Errorf("underlying ResponseWriter does not support hijacking")
+}
+
 // clientIP resolves the originating client address, preferring proxy headers
 // over the raw connection address.
 func clientIP(r *http.Request) string {

internal/server/log_test.go

@@ -1,11 +1,16 @@
 package server
 
 import (
+	"bufio"
 	"io"
+	"net"
 	"net/http"
 	"net/http/httptest"
+	"net/http/httputil"
+	"net/url"
 	"strings"
 	"testing"
+	"time"
 
 	"github.com/mostlygeek/llama-swap/internal/config"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
@@ -135,3 +140,103 @@ func TestServer_RequestLogMiddleware(t *testing.T) {
 		})
 	}
 }
+
+// TestServer_RequestLogMiddleware_WebSocketUpgrade verifies that the access-log
+// middleware (which wraps responses in statusRecorder) does not break websocket
+// upgrades proxied through httputil.ReverseProxy. ReverseProxy requires the
+// ResponseWriter to implement http.Hijacker to take over the connection; if
+// statusRecorder does not forward Hijack, the upgrade is refused with 502.
+func TestServer_RequestLogMiddleware_WebSocketUpgrade(t *testing.T) {
+	// Upstream: complete the upgrade handshake then echo bytes back. This
+	// stands in for an upstream that speaks websocket; ReverseProxy only cares
+	// about the 101 response and then copies raw bytes both ways.
+	upstream := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		hj, ok := w.(http.Hijacker)
+		if !ok {
+			t.Errorf("upstream ResponseWriter is not an http.Hijacker")
+			return
+		}
+		conn, brw, err := hj.Hijack()
+		if err != nil {
+			t.Errorf("upstream hijack: %v", err)
+			return
+		}
+		defer conn.Close()
+		brw.WriteString("HTTP/1.1 101 Switching Protocols\r\nUpgrade: websocket\r\nConnection: Upgrade\r\n\r\n")
+		brw.Flush()
+		// Echo whatever the client sends.
+		buf := make([]byte, 64)
+		n, err := brw.Read(buf)
+		if err != nil {
+			return
+		}
+		brw.Write(buf[:n])
+		brw.Flush()
+	}))
+	defer upstream.Close()
+
+	upstreamURL, err := url.Parse(upstream.URL)
+	if err != nil {
+		t.Fatalf("parse upstream URL: %v", err)
+	}
+
+	// Front server: ReverseProxy wrapped in the access-log middleware, which is
+	// the production statusRecorder-wrapped path.
+	proxy := httputil.NewSingleHostReverseProxy(upstreamURL)
+	mw := CreateRequestLogMiddleware(logmon.NewWriter(io.Discard))
+	front := httptest.NewServer(mw(proxy))
+	defer front.Close()
+
+	frontURL, err := url.Parse(front.URL)
+	if err != nil {
+		t.Fatalf("parse front URL: %v", err)
+	}
+
+	conn, err := net.DialTimeout("tcp", frontURL.Host, 5*time.Second)
+	if err != nil {
+		t.Fatalf("dial front: %v", err)
+	}
+	defer conn.Close()
+	conn.SetDeadline(time.Now().Add(5 * time.Second))
+
+	req := "GET / HTTP/1.1\r\n" +
+		"Host: " + frontURL.Host + "\r\n" +
+		"Connection: Upgrade\r\n" +
+		"Upgrade: websocket\r\n" +
+		"\r\n"
+	if _, err := conn.Write([]byte(req)); err != nil {
+		t.Fatalf("write upgrade request: %v", err)
+	}
+
+	br := bufio.NewReader(conn)
+	statusLine, err := br.ReadString('\n')
+	if err != nil {
+		t.Fatalf("read status line: %v", err)
+	}
+	if !strings.Contains(statusLine, "101") {
+		t.Fatalf("websocket upgrade failed: status line = %q, want 101 Switching Protocols", strings.TrimSpace(statusLine))
+	}
+
+	// Drain the rest of the response headers.
+	for {
+		line, err := br.ReadString('\n')
+		if err != nil {
+			t.Fatalf("read headers: %v", err)
+		}
+		if strings.TrimSpace(line) == "" {
+			break
+		}
+	}
+
+	// Verify bytes flow through the hijacked connection.
+	if _, err := conn.Write([]byte("ping")); err != nil {
+		t.Fatalf("write payload: %v", err)
+	}
+	echo := make([]byte, 4)
+	if _, err := io.ReadFull(br, echo); err != nil {
+		t.Fatalf("read echo: %v", err)
+	}
+	if string(echo) != "ping" {
+		t.Errorf("echo = %q, want %q", echo, "ping")
+	}
+}

internal/server/metrics.go

@@ -1,12 +1,14 @@
 package server
 
 import (
+	"bufio"
 	"bytes"
 	"compress/flate"
 	"compress/gzip"
 	"encoding/json"
 	"fmt"
 	"io"
+	"net"
 	"net/http"
 	"strings"
 	"sync"
@@ -23,6 +25,8 @@ import (
 // TokenMetrics holds token usage and performance metrics.
 type TokenMetrics struct {
 	CachedTokens    int     `json:"cache_tokens"`
+	DraftTokens     int     `json:"draft_tokens"`
+	DraftAccTokens  int     `json:"draft_acc_tokens"`
 	InputTokens     int     `json:"input_tokens"`
 	OutputTokens    int     `json:"output_tokens"`
 	PromptPerSecond float64 `json:"prompt_per_second"`
@@ -31,15 +35,17 @@ type TokenMetrics struct {
 
 // ActivityLogEntry represents parsed token statistics from llama-server logs.
 type ActivityLogEntry struct {
-	ID              int          `json:"id"`
-	Timestamp       time.Time    `json:"timestamp"`
-	Model           string       `json:"model"`
-	ReqPath         string       `json:"req_path"`
-	RespContentType string       `json:"resp_content_type"`
-	RespStatusCode  int          `json:"resp_status_code"`
-	Tokens          TokenMetrics `json:"tokens"`
-	DurationMs      int          `json:"duration_ms"`
-	HasCapture      bool         `json:"has_capture"`
+	ID              int               `json:"id"`
+	Timestamp       time.Time         `json:"timestamp"`
+	Model           string            `json:"model"`
+	ReqPath         string            `json:"req_path"`
+	RespContentType string            `json:"resp_content_type"`
+	RespStatusCode  int               `json:"resp_status_code"`
+	Tokens          TokenMetrics      `json:"tokens"`
+	DurationMs      int               `json:"duration_ms"`
+	HasCapture      bool              `json:"has_capture"`
+	ErrorMsg        string            `json:"error_msg,omitempty"`
+	Metadata        map[string]string `json:"metadata,omitempty"`
 }
 
 // ActivityLogEvent carries a single activity log entry to event subscribers.
@@ -120,9 +126,11 @@ func (mp *metricsMonitor) getMetricsJSON() ([]byte, error) {
 }
 
 // record parses a completed response body and stores/emits an activity entry.
-// When captures are enabled, a zstd+CBOR capture is stored for successful
-// requests, with cf controlling which request/response parts are retained.
-// reqBody and reqHeaders are the request data buffered before dispatch.
+// Successful requests store a zstd+CBOR capture (when enabled) with cf
+// controlling which parts are retained. Failed (non-200) requests capture the
+// request only and set ErrorMsg to a description of the failure, so the error
+// can be inspected without storing unreadable raw response bytes. reqBody and
+// reqHeaders are the request data buffered before dispatch.
 func (mp *metricsMonitor) record(modelID string, r *http.Request, recorder *responseBodyCopier, cf captureFields, reqBody []byte, reqHeaders map[string]string) {
 	tm := ActivityLogEntry{
 		Timestamp:       time.Now(),
@@ -133,6 +141,13 @@ func (mp *metricsMonitor) record(modelID string, r *http.Request, recorder *resp
 		DurationMs:      int(time.Since(recorder.StartTime()).Milliseconds()),
 	}
 
+	if ctxData, ok := shared.ReadContext(r.Context()); ok && len(ctxData.Metadata) > 0 {
+		tm.Metadata = make(map[string]string, len(ctxData.Metadata))
+		for k, v := range ctxData.Metadata {
+			tm.Metadata[k] = v
+		}
+	}
+
 	queueAndEmit := func() {
 		tm.ID = mp.queueMetrics(tm)
 		mp.emitMetric(tm)
@@ -140,7 +155,13 @@ func (mp *metricsMonitor) record(modelID string, r *http.Request, recorder *resp
 
 	if recorder.Status() != http.StatusOK {
 		mp.logger.Warnf("non-200 response, recording partial metrics: status=%d, path=%s", recorder.Status(), r.URL.Path)
-		queueAndEmit()
+		decoded, decErr := mp.decodeResponseBody(recorder, r.URL.Path)
+		tm.ErrorMsg = failedErrorMessage(recorder.Status(), decoded, decErr)
+		tm.ID = mp.queueMetrics(tm)
+		// Capture the request only; the failure is surfaced via ErrorMsg
+		// rather than storing the (possibly undisplayable) response body.
+		tm.HasCapture = mp.storeCapture(tm.ID, r, recorder, cf&^captureRespBody, reqBody, reqHeaders, nil)
+		mp.emitMetric(tm)
 		return
 	}
 
@@ -155,6 +176,7 @@ func (mp *metricsMonitor) record(modelID string, r *http.Request, recorder *resp
 		decoded, err := decompressBody(body, encoding)
 		if err != nil {
 			mp.logger.Warnf("metrics: decompression failed: %v, path=%s, recording minimal metrics", err, r.URL.Path)
+			tm.ErrorMsg = fmt.Sprintf("response decompression failed: %v", err)
 			queueAndEmit()
 			return
 		}
@@ -193,28 +215,99 @@ func (mp *metricsMonitor) record(modelID string, r *http.Request, recorder *resp
 	}
 
 	tm.ID = mp.queueMetrics(tm)
-	if mp.enableCaptures {
-		capture := ReqRespCapture{
-			ID:         tm.ID,
-			ReqPath:    r.URL.Path,
-			ReqHeaders: reqHeaders,
-		}
-		if cf&captureReqBody != 0 {
-			capture.ReqBody = reqBody
-		}
-		if cf&captureRespHeaders != 0 {
-			capture.RespHeaders = headerMap(recorder.Header())
-			redactHeaders(capture.RespHeaders)
-			delete(capture.RespHeaders, "Content-Encoding")
-		}
-		if cf&captureRespBody != 0 {
-			capture.RespBody = body
-		}
-		if mp.addCapture(capture) {
-			tm.HasCapture = true
+	tm.HasCapture = mp.storeCapture(tm.ID, r, recorder, cf, reqBody, reqHeaders, body)
+	mp.emitMetric(tm)
+}
+
+// storeCapture assembles a ReqRespCapture for id, honoring the captureFields
+// mask, and stores it when captures are enabled. body is the response body to
+// capture (already decompressed by the caller); pass nil to omit it. Returns
+// true if a capture was stored.
+func (mp *metricsMonitor) storeCapture(id int, r *http.Request, recorder *responseBodyCopier, cf captureFields, reqBody []byte, reqHeaders map[string]string, body []byte) bool {
+	if !mp.enableCaptures {
+		return false
+	}
+	capture := ReqRespCapture{
+		ID:         id,
+		ReqPath:    r.URL.Path,
+		ReqHeaders: reqHeaders,
+	}
+	if cf&captureReqBody != 0 {
+		capture.ReqBody = reqBody
+	}
+	if cf&captureRespHeaders != 0 {
+		capture.RespHeaders = headerMap(recorder.Header())
+		redactHeaders(capture.RespHeaders)
+		delete(capture.RespHeaders, "Content-Encoding")
+	}
+	if cf&captureRespBody != 0 {
+		capture.RespBody = body
+	}
+	return mp.addCapture(capture)
+}
+
+// decodeResponseBody returns the buffered response body, decompressing it when
+// the upstream set a Content-Encoding we recognize. On decompression failure it
+// logs a warning and returns an error so the caller can record a description
+// (via ErrorMsg) instead of storing unreadable raw bytes.
+func (mp *metricsMonitor) decodeResponseBody(recorder *responseBodyCopier, path string) ([]byte, error) {
+	body := recorder.body.Bytes()
+	if len(body) == 0 {
+		return nil, nil
+	}
+	encoding := recorder.Header().Get("Content-Encoding")
+	if encoding == "" {
+		return body, nil
+	}
+	decoded, err := decompressBody(body, encoding)
+	if err != nil {
+		mp.logger.Warnf("metrics: response decompression failed: %v, path=%s", err, path)
+		return nil, err
+	}
+	return decoded, nil
+}
+
+// errorMessagePaths lists JSON paths where a human-readable error message can
+// live across OpenAI- and llama.cpp-style error responses.
+var errorMessagePaths = []string{"error.message", "error", "message", "detail"}
+
+// extractErrorMessage pulls a human-readable error string from a JSON error
+// response. Returns "" if no message is found or the body is not valid JSON.
+func extractErrorMessage(body []byte) string {
+	if !gjson.ValidBytes(body) {
+		return ""
+	}
+	parsed := gjson.ParseBytes(body)
+	for _, path := range errorMessagePaths {
+		v := parsed.Get(path)
+		if v.Exists() && v.Type == gjson.String {
+			if s := strings.TrimSpace(v.String()); s != "" {
+				return s
+			}
 		}
 	}
-	mp.emitMetric(tm)
+	return ""
+}
+
+// failedErrorMessage builds a human-readable description for a non-200 response.
+// It prefers an error message parsed from the (decompressed) body and falls back
+// to the HTTP status text. A non-nil decErr indicates the body could not be
+// decoded, in which case the decode error is described instead.
+func failedErrorMessage(status int, body []byte, decErr error) string {
+	const maxLen = 500
+	if decErr != nil {
+		return fmt.Sprintf("response decode failed: %v", decErr)
+	}
+	if msg := extractErrorMessage(body); msg != "" {
+		if len(msg) > maxLen {
+			msg = msg[:maxLen] + "..."
+		}
+		return msg
+	}
+	if text := http.StatusText(status); text != "" {
+		return fmt.Sprintf("%d %s", status, text)
+	}
+	return fmt.Sprintf("HTTP %d", status)
 }
 
 // usagePaths lists the JSON paths where a per-event usage object can live.
@@ -335,6 +428,8 @@ func buildMetrics(modelID string, start time.Time, inputTokens, outputTokens, ca
 	durationMs := wallDurationMs
 	tokensPerSecond := -1.0
 	promptPerSecond := -1.0
+	draftTokens := -1
+	draftAccTokens := -1
 
 	if timings.Exists() {
 		inputTokens = timings.Get("prompt_n").Int()
@@ -348,6 +443,10 @@ func buildMetrics(modelID string, start time.Time, inputTokens, outputTokens, ca
 		if cachedValue := timings.Get("cache_n"); cachedValue.Exists() {
 			cachedTokens = cachedValue.Int()
 		}
+		if timings.Get("draft_n").Exists() && timings.Get("draft_n_accepted").Exists() {
+			draftTokens = int(timings.Get("draft_n").Int())
+			draftAccTokens = int(timings.Get("draft_n_accepted").Int())
+		}
 	}
 
 	return ActivityLogEntry{
@@ -355,6 +454,8 @@ func buildMetrics(modelID string, start time.Time, inputTokens, outputTokens, ca
 		Model:     modelID,
 		Tokens: TokenMetrics{
 			CachedTokens:    int(cachedTokens),
+			DraftTokens:     draftTokens,
+			DraftAccTokens:  draftAccTokens,
 			InputTokens:     int(inputTokens),
 			OutputTokens:    int(outputTokens),
 			PromptPerSecond: promptPerSecond,
@@ -427,6 +528,12 @@ func (w *responseBodyCopier) Write(b []byte) (int, error) {
 	if !w.wroteHeader {
 		w.WriteHeader(http.StatusOK)
 	}
+	// On a protocol upgrade (e.g. websocket) the body is raw framed data, not a
+	// metrics-parseable response, so write straight to the client without
+	// buffering a copy we can't use.
+	if w.status == http.StatusSwitchingProtocols {
+		return w.ResponseWriter.Write(b)
+	}
 	return w.tee.Write(b)
 }
 
@@ -446,5 +553,14 @@ func (w *responseBodyCopier) Flush() {
 	}
 }
 
+// Hijack forwards to the underlying writer so httputil.ReverseProxy can take
+// over the connection for websocket upgrades.
+func (w *responseBodyCopier) Hijack() (net.Conn, *bufio.ReadWriter, error) {
+	if hj, ok := w.ResponseWriter.(http.Hijacker); ok {
+		return hj.Hijack()
+	}
+	return nil, nil, fmt.Errorf("underlying ResponseWriter does not support hijacking")
+}
+
 func (w *responseBodyCopier) Status() int          { return w.status }
 func (w *responseBodyCopier) StartTime() time.Time { return w.start }

internal/server/metrics_middleware.go

@@ -4,10 +4,11 @@ import (
 	"bytes"
 	"io"
 	"net/http"
+	"strings"
 
 	"github.com/mostlygeek/llama-swap/internal/chain"
 	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/mostlygeek/llama-swap/internal/router"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
 // CreateMetricsMiddleware returns middleware that records token metrics for
@@ -21,17 +22,36 @@ func CreateMetricsMiddleware(mm *metricsMonitor, cfg config.Config) chain.Middle
 				return
 			}
 
+			// Determine the model-routed endpoint path. Regular routes are
+			// already meterable; /upstream/<model>/<path> is metered only when
+			// the remaining path matches a model-dispatched endpoint.
+			checkPath := r.URL.Path
+			if strings.HasPrefix(r.URL.Path, "/upstream/") {
+				var found bool
+				_, _, checkPath, found = shared.FindModelInPath(cfg, strings.TrimPrefix(r.URL.Path, "/upstream"))
+				if !found {
+					next.ServeHTTP(w, r)
+					return
+				}
+			}
+
+			if !isMetricsRecordPath(checkPath) {
+				next.ServeHTTP(w, r)
+				return
+			}
+
 			// Resolve the model now so downstream dispatch hits the context
-			// fast path; FetchContext restores the request body.
-			data, err := router.FetchContext(r, cfg)
+			// fast path; FetchContext restores the request body for regular
+			// routes and extracts the model from the URL for /upstream routes.
+			data, err := shared.FetchContext(r, cfg)
 			if err != nil {
-				router.SendError(w, r, router.ErrNoModelInContext)
+				shared.SendError(w, r, shared.ErrNoModelInContext)
 				return
 			}
 
 			// Buffer the request body/headers for capture before dispatch
 			// consumes them.
-			cf := captureFieldsFor(r.URL.Path)
+			cf := captureFieldsFor(checkPath)
 			var reqBody []byte
 			var reqHeaders map[string]string
 			if mm.enableCaptures {

internal/server/metrics_test.go

@@ -1,9 +1,16 @@
 package server
 
 import (
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"strings"
 	"testing"
 	"time"
 
+	"github.com/mostlygeek/llama-swap/internal/config"
+	"github.com/mostlygeek/llama-swap/internal/logmon"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 	"github.com/tidwall/gjson"
 )
 
@@ -56,6 +63,199 @@ func TestServer_ProcessStreamingResponse_NoData(t *testing.T) {
 	}
 }
 
+func TestMetricsMonitor_RecordMetadata(t *testing.T) {
+	mm := newMetricsMonitor(nil, 10, 0)
+	r := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(`{"usage":{}}`))
+	r = r.WithContext(shared.SetContext(r.Context(), shared.ReqContextData{
+		ModelID:  "m",
+		Metadata: map[string]string{"client": "web", "trace": "abc"},
+	}))
+
+	w := httptest.NewRecorder()
+	copier := newBodyCopier(w)
+	copier.WriteHeader(http.StatusOK)
+	copier.Write([]byte(`{"usage":{"prompt_tokens":1,"completion_tokens":2}}`))
+
+	mm.record("m", r, copier, 0, nil, nil)
+
+	entries := mm.getMetrics()
+	if len(entries) != 1 {
+		t.Fatalf("want 1 entry, got %d", len(entries))
+	}
+	if entries[0].Metadata["client"] != "web" {
+		t.Errorf("client = %q, want web", entries[0].Metadata["client"])
+	}
+	if entries[0].Metadata["trace"] != "abc" {
+		t.Errorf("trace = %q, want abc", entries[0].Metadata["trace"])
+	}
+}
+
+func TestMetricsMonitor_RecordFailedRequestCapture(t *testing.T) {
+	mm := newMetricsMonitor(logmon.NewWriter(io.Discard), 10, 5)
+	r := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", nil)
+	reqHeaders := map[string]string{"content-type": "application/json"}
+
+	w := httptest.NewRecorder()
+	copier := newBodyCopier(w)
+	copier.Header().Set("Content-Type", "application/json")
+	copier.WriteHeader(http.StatusBadGateway)
+	copier.Write([]byte(`{"error":{"message":"model unavailable"}}`))
+
+	reqBody := []byte(`{"model":"m","messages":[]}`)
+	mm.record("m", r, copier, captureAll, reqBody, reqHeaders)
+
+	entries := mm.getMetrics()
+	if len(entries) != 1 {
+		t.Fatalf("want 1 entry, got %d", len(entries))
+	}
+	entry := entries[0]
+	if entry.RespStatusCode != http.StatusBadGateway {
+		t.Errorf("status = %d, want %d", entry.RespStatusCode, http.StatusBadGateway)
+	}
+	if entry.ErrorMsg != "model unavailable" {
+		t.Errorf("error_msg = %q, want extracted message", entry.ErrorMsg)
+	}
+	if !entry.HasCapture {
+		t.Fatal("failed request should capture the request so it can be inspected")
+	}
+
+	got := mm.getCaptureByID(entry.ID)
+	if got == nil {
+		t.Fatal("capture not found")
+	}
+	if string(got.ReqBody) != `{"model":"m","messages":[]}` {
+		t.Errorf("req body = %q", got.ReqBody)
+	}
+	if len(got.RespBody) != 0 {
+		t.Errorf("resp body stored for failed request (len=%d); want none", len(got.RespBody))
+	}
+	if got.RespHeaders["Content-Type"] != "application/json" {
+		t.Errorf("resp Content-Type = %q", got.RespHeaders["Content-Type"])
+	}
+}
+
+func TestMetricsMonitor_RecordFailedRequestStatusFallback(t *testing.T) {
+	// Non-JSON error body: ErrorMsg falls back to the HTTP status text.
+	mm := newMetricsMonitor(logmon.NewWriter(io.Discard), 10, 5)
+	r := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", nil)
+
+	w := httptest.NewRecorder()
+	copier := newBodyCopier(w)
+	copier.WriteHeader(http.StatusBadGateway)
+	copier.Write([]byte("<html>upstream down</html>"))
+
+	mm.record("m", r, copier, captureAll, nil, nil)
+
+	entries := mm.getMetrics()
+	if len(entries) != 1 {
+		t.Fatalf("want 1 entry, got %d", len(entries))
+	}
+	if entries[0].ErrorMsg != "502 Bad Gateway" {
+		t.Errorf("error_msg = %q, want status text", entries[0].ErrorMsg)
+	}
+}
+
+func TestMetricsMonitor_RecordFailedRequestCaptureDisabled(t *testing.T) {
+	mm := newMetricsMonitor(logmon.NewWriter(io.Discard), 10, 0) // captures disabled
+	r := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", nil)
+
+	w := httptest.NewRecorder()
+	copier := newBodyCopier(w)
+	copier.WriteHeader(http.StatusInternalServerError)
+	copier.Write([]byte(`{"error":"boom"}`))
+
+	mm.record("m", r, copier, captureAll, []byte("req"), nil)
+
+	entries := mm.getMetrics()
+	if len(entries) != 1 {
+		t.Fatalf("want 1 entry, got %d", len(entries))
+	}
+	if entries[0].HasCapture {
+		t.Fatal("captures disabled, HasCapture should be false")
+	}
+	// ErrorMsg is independent of whether captures are enabled.
+	if entries[0].ErrorMsg != "boom" {
+		t.Errorf("error_msg = %q, want boom", entries[0].ErrorMsg)
+	}
+	if mm.getCaptureByID(entries[0].ID) != nil {
+		t.Fatal("no capture should be stored when disabled")
+	}
+}
+
+func TestMetricsMonitor_RecordDecompressionFailureSetsErrorMsg(t *testing.T) {
+	mm := newMetricsMonitor(logmon.NewWriter(io.Discard), 10, 5)
+	r := httptest.NewRequest(http.MethodPost, "/v1/chat/completions", nil)
+
+	w := httptest.NewRecorder()
+	copier := newBodyCopier(w)
+	copier.Header().Set("Content-Encoding", "gzip")
+	copier.WriteHeader(http.StatusOK)
+	copier.Write([]byte("not-really-gzip"))
+
+	mm.record("m", r, copier, captureAll, []byte("req"), nil)
+
+	entries := mm.getMetrics()
+	if len(entries) != 1 {
+		t.Fatalf("want 1 entry, got %d", len(entries))
+	}
+	if entries[0].ErrorMsg == "" {
+		t.Fatal("expected ErrorMsg for decompression failure")
+	}
+	// Raw bytes must not be stored when the body could not be decoded.
+	if entries[0].HasCapture {
+		t.Fatal("decompression failure should not store a capture")
+	}
+}
+
+func TestMetricsMonitor_DecodeResponseBody(t *testing.T) {
+	mm := newMetricsMonitor(logmon.NewWriter(io.Discard), 10, 5)
+
+	// No Content-Encoding: body returned unchanged.
+	w := httptest.NewRecorder()
+	copier := newBodyCopier(w)
+	copier.Write([]byte("plain"))
+	got, err := mm.decodeResponseBody(copier, "/p")
+	if err != nil || string(got) != "plain" {
+		t.Fatalf("plain body = %q, err = %v", got, err)
+	}
+
+	// Bogus gzip payload: returns an error and no body (no raw bytes kept).
+	w2 := httptest.NewRecorder()
+	copier2 := newBodyCopier(w2)
+	copier2.Header().Set("Content-Encoding", "gzip")
+	copier2.Write([]byte("not-really-gzip"))
+	got, err = mm.decodeResponseBody(copier2, "/p")
+	if err == nil {
+		t.Fatal("expected decompression error")
+	}
+	if got != nil {
+		t.Errorf("expected nil body on failure, got %q", got)
+	}
+}
+
+func TestServer_ExtractErrorMessage(t *testing.T) {
+	cases := []struct {
+		name string
+		body string
+		want string
+	}{
+		{"openai object", `{"error":{"message":"rate limited"}}`, "rate limited"},
+		{"string error", `{"error":"bad request"}`, "bad request"},
+		{"message field", `{"message":"nope"}`, "nope"},
+		{"detail field", `{"detail":"oops"}`, "oops"},
+		{"object error ignored", `{"error":{"code":42}}`, ""},
+		{"no error", `{"usage":{}}`, ""},
+		{"invalid json", `not-json`, ""},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			if got := extractErrorMessage([]byte(tc.body)); got != tc.want {
+				t.Errorf("extractErrorMessage = %q, want %q", got, tc.want)
+			}
+		})
+	}
+}
+
 func TestServer_ParseMetrics_Infill(t *testing.T) {
 	// /infill responses are arrays; timings live in the last element.
 	body := `[{"content":"a"},{"content":"b","timings":{"prompt_n":5,"predicted_n":9,"prompt_ms":10,"predicted_ms":20}}]`
@@ -72,3 +272,40 @@ func TestServer_ParseMetrics_Infill(t *testing.T) {
 		t.Fatalf("tokens = %+v", entry.Tokens)
 	}
 }
+
+// TestServer_MetricsMiddleware_UpstreamAudioCaptureSkipsRespBody verifies that
+// an /upstream/<model>/v1/audio/speech request uses the path-specific capture
+// mask (headers only) rather than falling back to captureAll.
+func TestServer_MetricsMiddleware_UpstreamAudioCaptureSkipsRespBody(t *testing.T) {
+	mm := newMetricsMonitor(logmon.NewWriter(io.Discard), 100, 5)
+	cfg := config.Config{Models: map[string]config.ModelConfig{"m1": {}}}
+
+	inner := http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.Header().Set("Content-Type", "audio/mpeg")
+		w.WriteHeader(http.StatusOK)
+		w.Write([]byte("BINARY-AUDIO-DATA"))
+	})
+	handler := CreateMetricsMiddleware(mm, cfg)(inner)
+
+	req := httptest.NewRequest(http.MethodPost, "/upstream/m1/v1/audio/speech", strings.NewReader(`{"model":"m1"}`))
+	handler.ServeHTTP(httptest.NewRecorder(), req)
+
+	entries := mm.getMetrics()
+	if len(entries) == 0 {
+		t.Fatal("no metrics recorded")
+	}
+	last := entries[len(entries)-1]
+	if !last.HasCapture {
+		t.Fatal("expected capture to be stored")
+	}
+	cap := mm.getCaptureByID(last.ID)
+	if cap == nil {
+		t.Fatal("capture not found")
+	}
+	if len(cap.RespBody) != 0 {
+		t.Errorf("RespBody stored for /upstream audio route (len=%d); want path-specific mask to skip body", len(cap.RespBody))
+	}
+	if len(cap.RespHeaders) == 0 {
+		t.Error("RespHeaders not stored; want captureRespHeaders mask")
+	}
+}

internal/server/server.go

@@ -15,6 +15,7 @@ import (
 	"github.com/mostlygeek/llama-swap/internal/logmon"
 	"github.com/mostlygeek/llama-swap/internal/perf"
 	"github.com/mostlygeek/llama-swap/internal/router"
+	"github.com/mostlygeek/llama-swap/internal/shared"
 )
 
 // Server owns the HTTP mux, cross-cutting middleware, and the local/peer model
@@ -88,6 +89,27 @@ var modelGetRoutes = []string{
 	"/sdapi/v1/loras",
 }
 
+// isMetricsRecordPath reports whether path is one of the model-dispatched
+// endpoints that the metrics middleware records in the activity log.
+func isMetricsRecordPath(path string) bool {
+	for _, p := range modelPostJSONRoutes {
+		if p == path {
+			return true
+		}
+	}
+	for _, p := range modelPostFormRoutes {
+		if p == path {
+			return true
+		}
+	}
+	for _, p := range modelGetRoutes {
+		if p == path {
+			return true
+		}
+	}
+	return false
+}
+
 // BuildInfo carries version metadata surfaced by GET /api/version.
 type BuildInfo struct {
 	Version string
@@ -99,12 +121,13 @@ func New(cfg config.Config, muxlog *logmon.Monitor, proxylog *logmon.Monitor, up
 	var local router.LocalRouter
 	var err error
 
-	if cfg.Matrix != nil {
+	switch cfg.Routing.Router.Use {
+	case "matrix":
 		local, err = router.NewMatrix(cfg, proxylog, upstreamlog)
 		if err != nil {
 			return nil, fmt.Errorf("creating matrix router: %w", err)
 		}
-	} else {
+	default: // "group"
 		local, err = router.NewGroup(cfg, proxylog, upstreamlog)
 		if err != nil {
 			return nil, fmt.Errorf("creating group router: %w", err)
@@ -137,13 +160,13 @@ func New(cfg config.Config, muxlog *logmon.Monitor, proxylog *logmon.Monitor, up
 }
 
 // localPeerHandler dispatches a model-routed request to the local or peer
-// router. The model is resolved once via router.FetchContext.
+// router. The model is resolved once via shared.FetchContext.
 func (s *Server) localPeerHandler(w http.ResponseWriter, r *http.Request) {
 	stripVersionPrefix(r)
 
-	data, err := router.FetchContext(r, s.cfg)
+	data, err := shared.FetchContext(r, s.cfg)
 	if err != nil {
-		router.SendError(w, r, router.ErrNoModelInContext)
+		shared.SendError(w, r, shared.ErrNoModelInContext)
 		return
 	}
 
@@ -155,7 +178,7 @@ func (s *Server) localPeerHandler(w http.ResponseWriter, r *http.Request) {
 		s.proxylog.Debugf("dispatch: using peer for model: %s", data.ModelID)
 		s.peer.ServeHTTP(w, r)
 	default:
-		router.SendError(w, r, router.ErrNoRouterFound)
+		shared.SendError(w, r, router.ErrNoRouterFound)
 	}
 }
 
@@ -170,21 +193,13 @@ func stripVersionPrefix(r *http.Request) {
 // routes builds the mux, registers every route, and wraps the mux with the
 // global CORS middleware.
 func (s *Server) routes() {
-	authMW := CreateAuthMiddleware(s.cfg)
-	filterMW := CreateFilterMiddleware(s.cfg)
-	formFilterMW := CreateFormFilterMiddleware(s.cfg)
 
-	// Model-dispatched routes get auth + per-model concurrency limiting + body
-	// filters + in-flight tracking + token metrics. concurrencyMW rejects with
-	// 429 before the body filters do any rewrite work. filterMW rewrites JSON
-	// bodies and formFilterMW rewrites multipart bodies; each is a no-op for the
-	// other's Content-Type. Both run before the metrics middleware so it buffers
-	// the rewritten body.
+	authMW := CreateAuthMiddleware(s.cfg)
 	modelChain := chain.New(
 		authMW,
-		CreateConcurrencyMiddleware(s.cfg),
-		filterMW,
-		formFilterMW,
+		CreateRequestContextMiddleware(s.cfg),
+		CreateFilterMiddleware(s.cfg),
+		CreateFormFilterMiddleware(s.cfg),
 		CreateInflightMiddleware(s.inflight),
 		CreateMetricsMiddleware(s.metrics, s.cfg),
 	)
@@ -215,19 +230,21 @@ func (s *Server) routes() {
 	mux.HandleFunc("GET /{$}", handleRootRedirect)
 
 	// Embedded UI.
-	mux.HandleFunc("GET /ui/", s.handleUI)
+	mux.Handle("GET /ui/", chain.New(authMW).ThenFunc(s.handleUI))
 	mux.HandleFunc("GET /favicon.ico", s.handleFavicon)
 
-	// Prometheus metrics (no auth, matches the legacy endpoint).
-	mux.HandleFunc("GET /metrics", s.handleMetrics)
+	// Prometheus metrics (wrapped by apiChain, matches the legacy endpoint).
+	mux.Handle("GET /metrics", apiChain.ThenFunc(s.handleMetrics))
 
 	// Operations endpoints.
 	mux.Handle("GET /unload", apiChain.ThenFunc(s.handleUnload))
 	mux.Handle("GET /running", apiChain.ThenFunc(s.handleRunning))
 
-	// Upstream passthrough.
+	// Upstream passthrough. Meter only the model-dispatched endpoints that can
+	// produce token usage/timings.
+	upstreamChain := apiChain.Append(CreateMetricsMiddleware(s.metrics, s.cfg))
 	mux.HandleFunc("GET /upstream", handleUpstreamRedirect)
-	mux.Handle("/upstream/{upstreamPath...}", apiChain.ThenFunc(s.handleUpstream))
+	mux.Handle("/upstream/{upstreamPath...}", upstreamChain.ThenFunc(s.handleUpstream))
 
 	// API group (API-key protected) consumed by the UI.
 	mux.Handle("POST /api/models/unload", apiChain.ThenFunc(s.handleAPIUnloadAll))

internal/server/server_test.go

@@ -84,10 +84,15 @@ func chatRequest(model string) *http.Request {
 
 func TestServer_New_GroupConfig(t *testing.T) {
 	discard := logmon.NewWriter(io.Discard)
-	s, err := New(config.Config{HealthCheckTimeout: 15}, discard, discard, discard, nil, BuildInfo{})
+	cfg := config.Config{HealthCheckTimeout: 15}
+	cfg.Routing.Router.Use = "group"
+	s, err := New(cfg, discard, discard, discard, nil, BuildInfo{})
 	if err != nil {
 		t.Fatalf("New (group): %v", err)
 	}
+	if _, ok := s.local.(*router.Group); !ok {
+		t.Fatalf("localRouter=%T want *router.Group", s.local)
+	}
 	if err := s.Shutdown(time.Second); err != nil {
 		t.Fatalf("Shutdown: %v", err)
 	}
@@ -95,11 +100,16 @@ func TestServer_New_GroupConfig(t *testing.T) {
 
 func TestServer_New_MatrixConfig(t *testing.T) {
 	discard := logmon.NewWriter(io.Discard)
-	cfg := config.Config{HealthCheckTimeout: 15, Matrix: &config.MatrixConfig{}}
+	cfg := config.Config{HealthCheckTimeout: 15}
+	cfg.Routing.Router.Use = "matrix"
+	cfg.Routing.Router.Settings.Matrix = &config.MatrixConfig{}
 	s, err := New(cfg, discard, discard, discard, nil, BuildInfo{})
 	if err != nil {
 		t.Fatalf("New (matrix): %v", err)
 	}
+	if _, ok := s.local.(*router.Matrix); !ok {
+		t.Fatalf("localRouter=%T want *router.Matrix", s.local)
+	}
 	if err := s.Shutdown(time.Second); err != nil {
 		t.Fatalf("Shutdown: %v", err)
 	}

internal/server/ui_dist/placeholder.txt

@@ -1 +0,0 @@
-placeholder so //go:embed ui_dist succeeds before the UI is built

internal/shared/http.go

@@ -0,0 +1,298 @@
+package shared
+
+import (
+	"bytes"
+	"context"
+	"encoding/base64"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"html"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/mostlygeek/llama-swap/internal/config"
+	"github.com/tidwall/gjson"
+)
+
+type contextkey struct {
+	name string
+}
+
+type ReqContextData struct {
+	ApiKey           string
+	Model            string
+	ModelID          string
+	Streaming        bool
+	SendLoadingState bool
+	// Metadata is a request-scoped key/value bag that handlers may mutate
+	// while processing. The metrics middleware copies it into ActivityLogEntry.
+	Metadata map[string]string
+}
+
+var (
+	ReqContextKey        = &contextkey{"context"}
+	ErrNoModelInContext  = fmt.Errorf("no model in request context")
+	ErrNoRouterFound     = fmt.Errorf("no router found for model")
+	ErrNoPeerModelFound  = fmt.Errorf("peer model not found")
+	ErrNoLocalModelFound = fmt.Errorf("local model not found")
+)
+
+func SendError(w http.ResponseWriter, r *http.Request, err error) {
+	var httpErr HTTPError
+	if errors.As(err, &httpErr) {
+		for k, v := range httpErr.Header() {
+			w.Header()[k] = v
+		}
+		w.WriteHeader(httpErr.StatusCode())
+		w.Write(httpErr.Body())
+		return
+	}
+
+	switch {
+	case errors.Is(err, ErrNoModelInContext):
+		SendResponse(w, r, http.StatusNotFound, "no model id could be identified")
+	case errors.Is(err, ErrNoPeerModelFound):
+		SendResponse(w, r, http.StatusNotFound, "no peer found for requested model")
+	case errors.Is(err, ErrNoLocalModelFound):
+		SendResponse(w, r, http.StatusNotFound, "no local server found for requested model")
+	case errors.Is(err, ErrNoRouterFound):
+		SendResponse(w, r, http.StatusNotFound, "no router for requested model")
+	default:
+		SendResponse(w, r, http.StatusInternalServerError, fmt.Sprintf("unspecific error: %v", err))
+	}
+}
+
+// SendResponse detects what content type the client prefers and returns an error response in that format.
+func SendResponse(w http.ResponseWriter, r *http.Request, status int, message string) {
+	acceptHeader := r.Header.Get("Accept")
+	if strings.Contains(acceptHeader, "text/plain") {
+		w.Header().Set("Content-Type", "text/plain")
+		w.WriteHeader(status)
+		w.Write([]byte(fmt.Sprintf("llama-swap: %s", message)))
+		return
+	}
+
+	if strings.Contains(acceptHeader, "text/html") {
+		w.Header().Set("Content-Type", "text/html")
+		w.WriteHeader(status)
+		w.Write([]byte(fmt.Sprintf(`<html><body><h1>llama-swap</h1><p>%s</p></body></html>`, html.EscapeString(message))))
+		return
+	}
+
+	w.Header().Set("Content-Type", "application/json")
+	w.WriteHeader(status)
+	resp, err := json.Marshal(map[string]string{"src": "llama-swap", "error": message})
+	if err != nil {
+		w.Write([]byte(`{"src":"llama-swap", "error": "failed to marshal response"}`))
+		return
+	}
+	w.Write(resp)
+}
+
+// FetchContext will attempt to get the model id from the context, then
+// from an /upstream/<model> path prefix, then from the request body/query.
+// If it extracts the model it will store it in the context for downstream
+// handlers. An error will be returned when a model cannot be identified.
+func FetchContext(r *http.Request, cfg config.Config) (ReqContextData, error) {
+	data, ok := ReadContext(r.Context())
+	if ok {
+		return data, nil
+	}
+
+	if strings.HasPrefix(r.URL.Path, "/upstream/") {
+		if data, ok := extractUpstreamContext(r, cfg); ok {
+			*r = *r.WithContext(SetContext(r.Context(), data))
+			return data, nil
+		}
+		return ReqContextData{}, ErrNoModelInContext
+	}
+
+	if data, err := extractContext(r); err == nil && data.Model != "" {
+		realName, _ := cfg.RealModelName(data.Model)
+		if realName == "" {
+			realName = data.Model
+		}
+		data.ModelID = realName
+		if mc, ok := cfg.Models[realName]; ok {
+			data.SendLoadingState = mc.SendLoadingState != nil && *mc.SendLoadingState
+		}
+		*r = *r.WithContext(SetContext(r.Context(), data))
+		return data, nil
+	}
+
+	return ReqContextData{}, ErrNoModelInContext
+}
+
+// extractUpstreamContext resolves the model from an /upstream/<model>/... path.
+func extractUpstreamContext(r *http.Request, cfg config.Config) (ReqContextData, bool) {
+	searchName, realName, _, found := FindModelInPath(cfg, strings.TrimPrefix(r.URL.Path, "/upstream"))
+	if !found {
+		return ReqContextData{}, false
+	}
+	return ReqContextData{
+		Model:            searchName,
+		ModelID:          realName,
+		ApiKey:           ExtractAPIKey(r),
+		Streaming:        r.URL.Query().Get("stream") == "true",
+		SendLoadingState: sendLoadingState(cfg, realName),
+		Metadata:         make(map[string]string),
+	}, true
+}
+
+// sendLoadingState reports whether the configured model wants loading-state SSEs.
+func sendLoadingState(cfg config.Config, modelID string) bool {
+	if mc, ok := cfg.Models[modelID]; ok {
+		return mc.SendLoadingState != nil && *mc.SendLoadingState
+	}
+	return false
+}
+
+// FindModelInPath walks a slash-separated path, building up segments until one
+// matches a configured model. This resolves model names that contain slashes
+// (e.g. "author/model"). Returns the matched name, its real model ID, the
+// remaining path, and whether a match was found.
+func FindModelInPath(cfg config.Config, path string) (searchName, realName, remainingPath string, found bool) {
+	parts := strings.Split(strings.TrimSpace(path), "/")
+	name := ""
+
+	for i, part := range parts {
+		if part == "" {
+			continue
+		}
+		if name == "" {
+			name = part
+		} else {
+			name = name + "/" + part
+		}
+
+		if modelID, ok := cfg.RealModelName(name); ok {
+			searchName = name
+			realName = modelID
+			remainingPath = "/" + strings.Join(parts[i+1:], "/")
+			found = true
+		}
+	}
+
+	return
+}
+
+func SetContext(ctx context.Context, data ReqContextData) context.Context {
+	return context.WithValue(ctx, ReqContextKey, data)
+}
+
+func ReadContext(ctx context.Context) (ReqContextData, bool) {
+	data, ok := ctx.Value(ReqContextKey).(ReqContextData)
+	return data, ok
+}
+
+// SetReqData attaches a key/value pair to the request context's metadata map.
+// The metadata map must already exist in the context's ReqContextData; callers
+// should ensure FetchContext has run or initialize the map themselves.
+// It returns an error for nil contexts or contexts without request data.
+func SetReqData(ctx context.Context, key, value string) error {
+	if ctx == nil {
+		return fmt.Errorf("cannot set request metadata on nil context")
+	}
+	data, ok := ReadContext(ctx)
+	if !ok {
+		return fmt.Errorf("no request context data found")
+	}
+	if data.Metadata == nil {
+		return fmt.Errorf("no metadata map in request context")
+	}
+	data.Metadata[key] = value
+	return nil
+}
+
+// extractContext pulls fields from an HTTP request into a ReqContextData,
+// returning whatever is available. For GET requests it reads query parameters.
+// For POST requests it inspects Content-Type and parses JSON,
+// multipart/form-data, or application/x-www-form-urlencoded bodies. The
+// request body is always restored before returning. An error is returned only
+// for I/O or parse failures, not for missing fields.
+func extractContext(r *http.Request) (ReqContextData, error) {
+
+	apiKey := ExtractAPIKey(r)
+
+	if r.Method == http.MethodGet {
+		q := r.URL.Query()
+		return ReqContextData{
+			Model:     q.Get("model"),
+			Streaming: q.Get("stream") == "true",
+			ApiKey:    apiKey,
+			Metadata:  make(map[string]string),
+		}, nil
+	}
+
+	bodyBytes, err := io.ReadAll(r.Body)
+	if err != nil {
+		return ReqContextData{}, fmt.Errorf("error reading request body: %w", err)
+	}
+	defer func() {
+		r.Body = io.NopCloser(bytes.NewReader(bodyBytes))
+	}()
+
+	contentType := r.Header.Get("Content-Type")
+
+	if strings.Contains(contentType, "application/json") {
+		return ReqContextData{
+			Model:     gjson.GetBytes(bodyBytes, "model").String(),
+			Streaming: gjson.GetBytes(bodyBytes, "stream").Bool(),
+			ApiKey:    apiKey,
+			Metadata:  make(map[string]string),
+		}, nil
+	}
+
+	// Form parsers read from r.Body, so feed them a fresh reader over the
+	// buffered bytes. The deferred restore above will reset r.Body again
+	// after parsing.
+	r.Body = io.NopCloser(bytes.NewReader(bodyBytes))
+	if strings.Contains(contentType, "multipart/form-data") {
+		if err := r.ParseMultipartForm(32 << 20); err != nil {
+			return ReqContextData{}, fmt.Errorf("error parsing multipart form: %w", err)
+		}
+	} else {
+		if err := r.ParseForm(); err != nil {
+			return ReqContextData{}, fmt.Errorf("error parsing form: %w", err)
+		}
+	}
+
+	return ReqContextData{
+		Model:     r.FormValue("model"),
+		Streaming: r.FormValue("stream") == "true",
+		ApiKey:    apiKey,
+		Metadata:  make(map[string]string),
+	}, nil
+}
+
+// extractAPIKey pulls a candidate API key from the request, preferring Basic,
+// then Bearer, then x-api-key.
+func ExtractAPIKey(r *http.Request) string {
+	var bearerKey, basicKey string
+	if auth := r.Header.Get("Authorization"); auth != "" {
+		scheme, credentials, ok := strings.Cut(auth, " ")
+		if ok {
+			switch strings.ToLower(scheme) {
+			case "bearer":
+				bearerKey = credentials
+			case "basic":
+				if decoded, err := base64.StdEncoding.DecodeString(credentials); err == nil {
+					if parts := strings.SplitN(string(decoded), ":", 2); len(parts) == 2 {
+						basicKey = parts[1] // password field is the API key
+					}
+				}
+			}
+		}
+	}
+
+	switch {
+	case basicKey != "":
+		return basicKey
+	case bearerKey != "":
+		return bearerKey
+	default:
+		return r.Header.Get("x-api-key")
+	}
+}

internal/shared/http_test.go

@@ -0,0 +1,525 @@
+package shared
+
+import (
+	"bytes"
+	"context"
+	"encoding/base64"
+	"io"
+	"mime/multipart"
+	"net/http"
+	"net/http/httptest"
+	"net/url"
+	"strings"
+	"testing"
+
+	"github.com/mostlygeek/llama-swap/internal/config"
+)
+
+func TestExtractContext_GET(t *testing.T) {
+	tests := []struct {
+		name      string
+		query     string
+		wantModel string
+		wantErr   bool
+	}{
+		{"model present", "model=llama3", "llama3", false},
+		{"model with slashes", "model=author/model-7b", "author/model-7b", false},
+		{"model missing", "", "", false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			r, _ := http.NewRequest(http.MethodGet, "/?"+tt.query, nil)
+			got, err := extractContext(r)
+			if (err != nil) != tt.wantErr {
+				t.Fatalf("wantErr=%v got err=%v", tt.wantErr, err)
+			}
+			if got.Model != tt.wantModel {
+				t.Errorf("want %q got %q", tt.wantModel, got.Model)
+			}
+		})
+	}
+}
+
+func TestExtractContext_JSON(t *testing.T) {
+	tests := []struct {
+		name      string
+		body      string
+		wantModel string
+		wantErr   bool
+	}{
+		{"model present", `{"model":"llama3","stream":true}`, "llama3", false},
+		{"model with slashes", `{"model":"author/model-7b"}`, "author/model-7b", false},
+		{"model empty string", `{"model":""}`, "", false},
+		{"model key missing", `{"stream":true}`, "", false},
+		{"invalid json", `not-json`, "", false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			r, _ := http.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(tt.body))
+			r.Header.Set("Content-Type", "application/json")
+			got, err := extractContext(r)
+			if (err != nil) != tt.wantErr {
+				t.Fatalf("wantErr=%v got err=%v", tt.wantErr, err)
+			}
+			if got.Model != tt.wantModel {
+				t.Errorf("want %q got %q", tt.wantModel, got.Model)
+			}
+		})
+	}
+}
+
+func TestExtractContext_URLEncodedForm(t *testing.T) {
+	tests := []struct {
+		name      string
+		formModel string
+		wantModel string
+		wantErr   bool
+	}{
+		{"model present", "whisper-1", "whisper-1", false},
+		{"model missing", "", "", false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			form := url.Values{}
+			if tt.formModel != "" {
+				form.Set("model", tt.formModel)
+			}
+			r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", strings.NewReader(form.Encode()))
+			r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+			got, err := extractContext(r)
+			if (err != nil) != tt.wantErr {
+				t.Fatalf("wantErr=%v got err=%v", tt.wantErr, err)
+			}
+			if got.Model != tt.wantModel {
+				t.Errorf("want %q got %q", tt.wantModel, got.Model)
+			}
+		})
+	}
+}
+
+func TestExtractContext_MultipartForm(t *testing.T) {
+	tests := []struct {
+		name      string
+		formModel string
+		wantModel string
+		wantErr   bool
+	}{
+		{"model present", "whisper-1", "whisper-1", false},
+		{"model missing", "", "", false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			var buf bytes.Buffer
+			mw := multipart.NewWriter(&buf)
+			if tt.formModel != "" {
+				fw, _ := mw.CreateFormField("model")
+				fw.Write([]byte(tt.formModel))
+			}
+			mw.Close()
+
+			r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", &buf)
+			r.Header.Set("Content-Type", mw.FormDataContentType())
+			got, err := extractContext(r)
+			if (err != nil) != tt.wantErr {
+				t.Fatalf("wantErr=%v got err=%v", tt.wantErr, err)
+			}
+			if got.Model != tt.wantModel {
+				t.Errorf("want %q got %q", tt.wantModel, got.Model)
+			}
+		})
+	}
+}
+
+func TestExtractContext_JSONBodyRestored(t *testing.T) {
+	body := `{"model":"llama3","stream":true}`
+	r, _ := http.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(body))
+	r.Header.Set("Content-Type", "application/json")
+
+	if _, err := extractContext(r); err != nil {
+		t.Fatalf("ExtractContext: %v", err)
+	}
+
+	remaining, err := io.ReadAll(r.Body)
+	if err != nil {
+		t.Fatalf("reading body after ExtractContext: %v", err)
+	}
+	if string(remaining) != body {
+		t.Errorf("body not restored: want %q got %q", body, string(remaining))
+	}
+}
+
+func TestExtractContext_MultipartBodyRestored(t *testing.T) {
+	var buf bytes.Buffer
+	mw := multipart.NewWriter(&buf)
+	fw, _ := mw.CreateFormField("model")
+	fw.Write([]byte("whisper-1"))
+	ff, _ := mw.CreateFormFile("file", "audio.wav")
+	ff.Write([]byte("fake-audio-bytes"))
+	mw.Close()
+
+	original := buf.Bytes()
+
+	r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", bytes.NewReader(original))
+	r.Header.Set("Content-Type", mw.FormDataContentType())
+
+	if _, err := extractContext(r); err != nil {
+		t.Fatalf("ExtractContext: %v", err)
+	}
+
+	remaining, err := io.ReadAll(r.Body)
+	if err != nil {
+		t.Fatalf("reading body after ExtractContext: %v", err)
+	}
+	if !bytes.Equal(remaining, original) {
+		t.Errorf("multipart body not restored: want %d bytes got %d bytes", len(original), len(remaining))
+	}
+}
+
+func TestExtractContext_URLEncodedBodyRestored(t *testing.T) {
+	body := "model=whisper-1&extra=value"
+	r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", strings.NewReader(body))
+	r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+
+	if _, err := extractContext(r); err != nil {
+		t.Fatalf("ExtractContext: %v", err)
+	}
+
+	remaining, err := io.ReadAll(r.Body)
+	if err != nil {
+		t.Fatalf("reading body after ExtractContext: %v", err)
+	}
+	if string(remaining) != body {
+		t.Errorf("url-encoded body not restored: want %q got %q", body, string(remaining))
+	}
+}
+
+func TestSetContext(t *testing.T) {
+	ctx := SetContext(context.Background(), ReqContextData{Model: "llama3", ModelID: "llama3"})
+	data, ok := ctx.Value(ReqContextKey).(ReqContextData)
+	if !ok {
+		t.Fatalf("ContextKey not set or wrong type")
+	}
+	if data.Model != "llama3" {
+		t.Errorf("want %q got %q", "llama3", data.Model)
+	}
+	if data.ModelID != "llama3" {
+		t.Errorf("want %q got %q", "llama3", data.ModelID)
+	}
+}
+
+func TestSetContext_WithAlias(t *testing.T) {
+	ctx := SetContext(context.Background(), ReqContextData{Model: "llama", ModelID: "llama3"})
+	data, _ := ctx.Value(ReqContextKey).(ReqContextData)
+	if data.Model != "llama" {
+		t.Errorf("want requested %q got %q", "llama", data.Model)
+	}
+	if data.ModelID != "llama3" {
+		t.Errorf("want real %q got %q", "llama3", data.ModelID)
+	}
+}
+
+func TestSetContext_DoesNotMutateParent(t *testing.T) {
+	parent := context.Background()
+	_ = SetContext(parent, ReqContextData{Model: "llama3", ModelID: "llama3"})
+	if v := parent.Value(ReqContextKey); v != nil {
+		t.Errorf("parent context was mutated: %v", v)
+	}
+}
+
+func TestReadContext(t *testing.T) {
+	tests := []struct {
+		name     string
+		ctx      context.Context
+		wantReq  string
+		wantReal string
+		wantBool bool
+	}{
+		{
+			name:     "model present, same name",
+			ctx:      SetContext(context.Background(), ReqContextData{Model: "llama3", ModelID: "llama3"}),
+			wantReq:  "llama3",
+			wantReal: "llama3",
+			wantBool: true,
+		},
+		{
+			name:     "model present, aliased",
+			ctx:      SetContext(context.Background(), ReqContextData{Model: "llama", ModelID: "llama3"}),
+			wantReq:  "llama",
+			wantReal: "llama3",
+			wantBool: true,
+		},
+		{
+			name:     "model absent",
+			ctx:      context.Background(),
+			wantReq:  "",
+			wantReal: "",
+			wantBool: false,
+		},
+		{
+			name:     "model is empty string",
+			ctx:      SetContext(context.Background(), ReqContextData{Model: "", ModelID: ""}),
+			wantReq:  "",
+			wantReal: "",
+			wantBool: true,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			gotData, ok := ReadContext(tt.ctx)
+			if gotData.Model != tt.wantReq || gotData.ModelID != tt.wantReal || ok != tt.wantBool {
+				t.Errorf("want (%q, %q, %v) got (%q, %q, %v)", tt.wantReq, tt.wantReal, tt.wantBool, gotData.Model, gotData.ModelID, ok)
+			}
+		})
+	}
+}
+
+func TestExtractContext_Streaming_GET(t *testing.T) {
+	tests := []struct {
+		name          string
+		query         string
+		wantStreaming bool
+	}{
+		{"streaming true", "model=llama3&stream=true", true},
+		{"streaming false", "model=llama3&stream=false", false},
+		{"no stream param", "model=llama3", false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			r, _ := http.NewRequest(http.MethodGet, "/?"+tt.query, nil)
+			got, err := extractContext(r)
+			if err != nil {
+				t.Fatalf("ExtractContext: %v", err)
+			}
+			if got.Streaming != tt.wantStreaming {
+				t.Errorf("Streaming: want %v, got %v", tt.wantStreaming, got.Streaming)
+			}
+		})
+	}
+}
+
+func TestExtractContext_Streaming_JSON(t *testing.T) {
+	tests := []struct {
+		name          string
+		body          string
+		wantStreaming bool
+	}{
+		{"streaming true", `{"model":"llama3","stream":true}`, true},
+		{"streaming false", `{"model":"llama3","stream":false}`, false},
+		{"no stream param", `{"model":"llama3"}`, false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			r, _ := http.NewRequest(http.MethodPost, "/v1/chat/completions", strings.NewReader(tt.body))
+			r.Header.Set("Content-Type", "application/json")
+			got, err := extractContext(r)
+			if err != nil {
+				t.Fatalf("ExtractContext: %v", err)
+			}
+			if got.Streaming != tt.wantStreaming {
+				t.Errorf("Streaming: want %v, got %v", tt.wantStreaming, got.Streaming)
+			}
+		})
+	}
+}
+
+func TestExtractContext_Streaming_URLEncodedForm(t *testing.T) {
+	r, _ := http.NewRequest(http.MethodPost, "/v1/audio/transcriptions", strings.NewReader("model=whisper-1&stream=true"))
+	r.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	got, err := extractContext(r)
+	if err != nil {
+		t.Fatalf("ExtractContext: %v", err)
+	}
+	if !got.Streaming {
+		t.Error("Streaming should be true")
+	}
+}
+
+func TestExtractContext_ApiKey(t *testing.T) {
+	basicHeader := func(user, pass string) string {
+		return "Basic " + base64.StdEncoding.EncodeToString([]byte(user+":"+pass))
+	}
+	cases := []struct {
+		name    string
+		method  string
+		ct      string
+		body    string
+		auth    string
+		xapi    string
+		wantKey string
+	}{
+		{"GET bearer", http.MethodGet, "", "", "Bearer sk-get", "", "sk-get"},
+		{"GET x-api-key", http.MethodGet, "", "", "", "xk-get", "xk-get"},
+		{"GET basic", http.MethodGet, "", "", basicHeader("u", "pw-get"), "", "pw-get"},
+		{"JSON bearer", http.MethodPost, "application/json", `{"model":"m"}`, "Bearer sk-json", "", "sk-json"},
+		{"JSON x-api-key", http.MethodPost, "application/json", `{"model":"m"}`, "", "xk-json", "xk-json"},
+		{"form bearer", http.MethodPost, "application/x-www-form-urlencoded", "model=m", "Bearer sk-form", "", "sk-form"},
+		{"no key", http.MethodGet, "", "", "", "", ""},
+	}
+	for _, c := range cases {
+		t.Run(c.name, func(t *testing.T) {
+			var body io.Reader
+			if c.body != "" {
+				body = strings.NewReader(c.body)
+			}
+			r, _ := http.NewRequest(c.method, "/", body)
+			if c.ct != "" {
+				r.Header.Set("Content-Type", c.ct)
+			}
+			if c.auth != "" {
+				r.Header.Set("Authorization", c.auth)
+			}
+			if c.xapi != "" {
+				r.Header.Set("x-api-key", c.xapi)
+			}
+			got, err := extractContext(r)
+			if err != nil {
+				t.Fatalf("extractContext: %v", err)
+			}
+			if got.ApiKey != c.wantKey {
+				t.Errorf("ApiKey = %q, want %q", got.ApiKey, c.wantKey)
+			}
+		})
+	}
+}
+
+func TestSetReqData(t *testing.T) {
+	ctx := SetContext(context.Background(), ReqContextData{Model: "llama3", ModelID: "llama3", Metadata: make(map[string]string)})
+
+	if err := SetReqData(ctx, "client", "web"); err != nil {
+		t.Fatalf("SetReqData: %v", err)
+	}
+	if err := SetReqData(ctx, "trace", "abc123"); err != nil {
+		t.Fatalf("SetReqData: %v", err)
+	}
+
+	data, ok := ReadContext(ctx)
+	if !ok {
+		t.Fatal("context data missing")
+	}
+	if data.Metadata["client"] != "web" {
+		t.Errorf("client = %q, want %q", data.Metadata["client"], "web")
+	}
+	if data.Metadata["trace"] != "abc123" {
+		t.Errorf("trace = %q, want %q", data.Metadata["trace"], "abc123")
+	}
+}
+
+func TestSetReqData_Errors(t *testing.T) {
+	if err := SetReqData(context.Background(), "k", "v"); err == nil {
+		t.Error("expected error when no request context data exists")
+	}
+	ctx := SetContext(context.Background(), ReqContextData{Model: "llama3", ModelID: "llama3"})
+	if err := SetReqData(ctx, "k", "v"); err == nil {
+		t.Error("expected error when metadata map is missing")
+	}
+}
+
+func TestServer_ExtractAPIKey(t *testing.T) {
+	basicHeader := func(user, pass string) string {
+		return "Basic " + base64.StdEncoding.EncodeToString([]byte(user+":"+pass))
+	}
+	cases := []struct {
+		name string
+		auth string
+		xapi string
+		want string
+	}{
+		{"none", "", "", ""},
+		{"bearer", "Bearer tok123", "", "tok123"},
+		{"basic", basicHeader("user", "pw-key"), "", "pw-key"},
+		{"x-api-key", "", "xkey", "xkey"},
+		{"basic beats bearer", basicHeader("u", "bk"), "", "bk"},
+		{"bearer beats x-api-key", "Bearer btok", "xkey", "btok"},
+		{"malformed basic falls back to x-api-key", "Basic !!!notbase64", "xkey", "xkey"},
+		{"lowercase bearer", "bearer tok123", "", "tok123"},
+		{"lowercase basic", "basic " + base64.StdEncoding.EncodeToString([]byte("user:pw-key")), "", "pw-key"},
+		{"mixed case BEARER", "BEARER tok456", "", "tok456"},
+		{"mixed case bAsIc", "bAsIc " + base64.StdEncoding.EncodeToString([]byte("u:bk")), "", "bk"},
+	}
+	for _, c := range cases {
+		t.Run(c.name, func(t *testing.T) {
+			r := httptest.NewRequest(http.MethodGet, "/", nil)
+			if c.auth != "" {
+				r.Header.Set("Authorization", c.auth)
+			}
+			if c.xapi != "" {
+				r.Header.Set("x-api-key", c.xapi)
+			}
+			if got := ExtractAPIKey(r); got != c.want {
+				t.Errorf("extractAPIKey() = %q, want %q", got, c.want)
+			}
+		})
+	}
+}
+
+func TestFetchContext_UpstreamPath(t *testing.T) {
+	cfg := config.Config{
+		Models: map[string]config.ModelConfig{
+			"m1":           {},
+			"author/model": {},
+			"real":         {Aliases: []string{"nick"}},
+		},
+	}
+
+	cases := []struct {
+		name        string
+		path        string
+		wantModel   string
+		wantModelID string
+		wantErr     bool
+	}{
+		{"known model", "/upstream/m1/v1/chat/completions", "m1", "m1", false},
+		{"model with slash", "/upstream/author/model/v1/chat", "author/model", "author/model", false},
+		{"unknown model", "/upstream/nope/v1/chat/completions", "", "", true},
+		{"bare model path", "/upstream/m1/", "m1", "m1", false},
+	}
+
+	for _, c := range cases {
+		t.Run(c.name, func(t *testing.T) {
+			r := httptest.NewRequest(http.MethodPost, c.path, strings.NewReader(`{}`))
+			data, err := FetchContext(r, cfg)
+			if (err != nil) != c.wantErr {
+				t.Fatalf("wantErr=%v got err=%v", c.wantErr, err)
+			}
+			if c.wantErr {
+				return
+			}
+			if data.Model != c.wantModel {
+				t.Errorf("model = %q, want %q", data.Model, c.wantModel)
+			}
+			if data.ModelID != c.wantModelID {
+				t.Errorf("modelID = %q, want %q", data.ModelID, c.wantModelID)
+			}
+			if data.Metadata == nil {
+				t.Error("metadata map not initialized")
+			}
+		})
+	}
+}
+
+func TestFetchContext_UpstreamPath_DoesNotReadBody(t *testing.T) {
+	cfg := config.Config{Models: map[string]config.ModelConfig{"m1": {}}}
+	body := `{"model":"should-not-matter"}`
+	r := httptest.NewRequest(http.MethodPost, "/upstream/m1/v1/chat/completions", strings.NewReader(body))
+
+	_, err := FetchContext(r, cfg)
+	if err != nil {
+		t.Fatalf("FetchContext: %v", err)
+	}
+
+	// The body should be untouched so the upstream handler can still read it.
+	got, err := io.ReadAll(r.Body)
+	if err != nil {
+		t.Fatalf("read body: %v", err)
+	}
+	if string(got) != body {
+		t.Errorf("body was consumed: %q", string(got))
+	}
+}

internal/shared/httperror.go

@@ -0,0 +1,63 @@
+package shared
+
+import (
+	"encoding/json"
+	"net/http"
+	"strconv"
+)
+
+// HTTPError is an error that carries a complete HTTP response. A producer (e.g.
+// a scheduler shedding a request) returns one of these; a renderer (e.g.
+// router.SendError) writes the status, headers, and body verbatim instead of
+// mapping the error to a generic status. It is the seam that lets a component
+// shed a request with a rich response (e.g. a 429 with rate-limit headers and a
+// JSON hint body) without the renderer knowing the producer's internals.
+type HTTPError interface {
+	error
+	StatusCode() int
+	Header() http.Header
+	Body() []byte
+}
+
+// ConcurrencyLimitError is an HTTPError for a 429 concurrency-limit rejection.
+// Zero-value fields fall back to sensible defaults: a 1-second Retry-After and a
+// JSON hint body.
+type ConcurrencyLimitError struct {
+	// RetryAfter, when > 0, is sent as the Retry-After header (in seconds).
+	// Defaults to 1.
+	RetryAfter int
+
+	// Message overrides the JSON body's "error" field. Defaults to
+	// "Too many requests".
+	Message string
+}
+
+func (e ConcurrencyLimitError) Error() string { return "concurrency limit reached" }
+
+func (e ConcurrencyLimitError) StatusCode() int { return http.StatusTooManyRequests }
+
+func (e ConcurrencyLimitError) Header() http.Header {
+	h := http.Header{}
+	h.Set("Content-Type", "application/json")
+	h.Set("Retry-After", e.retryAfter())
+	return h
+}
+
+func (e ConcurrencyLimitError) Body() []byte {
+	b, _ := json.Marshal(map[string]string{"error": e.message()})
+	return b
+}
+
+func (e ConcurrencyLimitError) retryAfter() string {
+	if e.RetryAfter > 0 {
+		return strconv.Itoa(e.RetryAfter)
+	}
+	return "1"
+}
+
+func (e ConcurrencyLimitError) message() string {
+	if e.Message != "" {
+		return e.Message
+	}
+	return "Too many requests"
+}

internal/shared/loopback.go

@@ -0,0 +1,31 @@
+package shared
+
+import "net"
+
+// IsLoopbackAddr reports whether listenAddr binds exclusively to loopback.
+// Addresses with an empty or wildcard host (e.g. ":8080", "0.0.0.0:8080",
+// "[::]:8080") bind on all interfaces and return false.
+func IsLoopbackAddr(listenAddr string) bool {
+	host, _, err := net.SplitHostPort(listenAddr)
+	if err != nil {
+		return false
+	}
+	if host == "" {
+		return false
+	}
+	ip := net.ParseIP(host)
+	if ip != nil {
+		return ip.IsLoopback()
+	}
+	// hostname case (e.g. "localhost")
+	addrs, err := net.LookupHost(host)
+	if err != nil {
+		return false
+	}
+	for _, a := range addrs {
+		if !net.ParseIP(a).IsLoopback() {
+			return false
+		}
+	}
+	return len(addrs) > 0
+}

internal/watcher/dirwatcher.go

@@ -0,0 +1,137 @@
+package configwatcher
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+	"sort"
+	"strings"
+	"time"
+)
+
+// DirWatcher polls a directory for changes to its set of *.yml / *.yaml files.
+// It fires OnChange when a file is added, removed, or has its mod time/size
+// change. Like Watcher it is poll-based so it works in Docker bind-mounts and
+// k8s ConfigMap projections where inotify is unreliable.
+//
+// The baseline poll establishes initial state and does not fire OnChange.
+type DirWatcher struct {
+	Path     string
+	Interval time.Duration
+	OnChange func()
+}
+
+// dirSnapshot is an ordered map of file name -> file state. The ordering is
+// derived from sorted filenames so two snapshots compare deterministically
+// regardless of readdir order. exists reflects whether the directory was
+// readable at scan time; a missing directory yields exists=false.
+type dirSnapshot struct {
+	exists bool
+	names  []string
+	states map[string]snapshot
+}
+
+func newDirSnapshot() dirSnapshot {
+	return dirSnapshot{states: make(map[string]snapshot)}
+}
+
+// equal reports whether two snapshots describe the same file set and per-file
+// state. A missing directory (exists=false) is treated as equal to any other
+// missing directory regardless of cached names.
+func (s dirSnapshot) equal(other dirSnapshot) bool {
+	if !s.exists && !other.exists {
+		return true
+	}
+	if s.exists != other.exists {
+		return false
+	}
+	if len(s.names) != len(other.names) {
+		return false
+	}
+	for i, n := range s.names {
+		if other.names[i] != n {
+			return false
+		}
+	}
+	for _, n := range s.names {
+		a, b := s.states[n], other.states[n]
+		if a.exists != b.exists || a.size != b.size || !a.modTime.Equal(b.modTime) {
+			return false
+		}
+	}
+	return true
+}
+
+// Run blocks until ctx is canceled. It polls Path on Interval and invokes
+// OnChange whenever the directory's YAML file set changes.
+//
+// Policy mirrors the single-file Watcher: disappearance (directory missing or
+// empty) is treated as a transient rename-style write and stays quiet; the
+// transition back to present-with-content fires OnChange.
+func (w *DirWatcher) Run(ctx context.Context) {
+	interval := w.Interval
+	if interval <= 0 {
+		interval = DefaultInterval
+	}
+
+	prev := scanDir(w.Path)
+
+	ticker := time.NewTicker(interval)
+	defer ticker.Stop()
+
+	for {
+		select {
+		case <-ctx.Done():
+			return
+		case <-ticker.C:
+			cur := scanDir(w.Path)
+			// Suppress transitions involving an empty or missing directory —
+			// these are treated as transient rename-style writes, mirroring
+			// the single-file Watcher. Only present-with-content →
+			// present-with-content (changed) or no-content →
+			// present-with-content fires OnChange.
+			prevHasContent := prev.exists && len(prev.names) > 0
+			curHasContent := cur.exists && len(cur.names) > 0
+			if curHasContent && (!prevHasContent || !prev.equal(cur)) && w.OnChange != nil {
+				w.OnChange()
+			}
+			prev = cur
+		}
+	}
+}
+
+// scanDir returns a snapshot of the *.yml/*.yaml files in dir. If the
+// directory cannot be read (missing, permission denied) the snapshot reports
+// exists=false; the next successful scan will detect the recovery and fire
+// OnChange.
+func scanDir(dir string) dirSnapshot {
+	snap := newDirSnapshot()
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		return snap // exists=false
+	}
+	snap.exists = true
+	for _, e := range entries {
+		if e.IsDir() {
+			continue
+		}
+		name := e.Name()
+		if !strings.HasSuffix(name, ".yml") && !strings.HasSuffix(name, ".yaml") {
+			continue
+		}
+		fi, err := os.Stat(filepath.Join(dir, name))
+		if err != nil {
+			// File disappeared between ReadDir and Stat; skip it — the
+			// next poll will observe the removal cleanly.
+			continue
+		}
+		snap.names = append(snap.names, name)
+		snap.states[name] = snapshot{
+			exists:  true,
+			modTime: fi.ModTime(),
+			size:    fi.Size(),
+		}
+	}
+	sort.Strings(snap.names)
+	return snap
+}

internal/watcher/dirwatcher_test.go

@@ -0,0 +1,199 @@
+package configwatcher
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+)
+
+// startDirWatcher launches w.Run in a goroutine and returns a function that
+// cancels the context and waits for Run to return.
+func startDirWatcher(t *testing.T, w *DirWatcher) func() {
+	t.Helper()
+	ctx, cancel := context.WithCancel(context.Background())
+	done := make(chan struct{})
+	go func() {
+		w.Run(ctx)
+		close(done)
+	}()
+	return func() {
+		cancel()
+		select {
+		case <-done:
+		case <-time.After(2 * time.Second):
+			t.Fatal("DirWatcher did not stop within 2s of cancel")
+		}
+	}
+}
+
+func writeYAMLInDir(t *testing.T, dir, name, content string) {
+	t.Helper()
+	require.NoError(t, os.WriteFile(filepath.Join(dir, name), []byte(content), 0o644))
+}
+
+func TestDirWatcher_NoFireOnBaseline(t *testing.T) {
+	dir := t.TempDir()
+	writeYAMLInDir(t, dir, "a.yaml", "a")
+
+	var n int64
+	stop := startDirWatcher(t, &DirWatcher{
+		Path:     dir,
+		Interval: testInterval,
+		OnChange: func() { atomic.AddInt64(&n, 1) },
+	})
+	defer stop()
+
+	time.Sleep(testInterval * 5)
+	require.Equal(t, int64(0), atomic.LoadInt64(&n), "baseline poll must not fire")
+}
+
+func TestDirWatcher_DetectsFileAdd(t *testing.T) {
+	dir := t.TempDir()
+	writeYAMLInDir(t, dir, "a.yaml", "a")
+
+	var n int64
+	stop := startDirWatcher(t, &DirWatcher{
+		Path:     dir,
+		Interval: testInterval,
+		OnChange: func() { atomic.AddInt64(&n, 1) },
+	})
+	defer stop()
+	time.Sleep(testInterval * 2)
+
+	writeYAMLInDir(t, dir, "b.yaml", "b")
+	require.True(t, waitForCount(t, &n, 1, time.Second), "callback should fire when a file is added")
+}
+
+func TestDirWatcher_DetectsFileRemoval(t *testing.T) {
+	dir := t.TempDir()
+	writeYAMLInDir(t, dir, "a.yaml", "a")
+	writeYAMLInDir(t, dir, "b.yaml", "b")
+
+	var n int64
+	stop := startDirWatcher(t, &DirWatcher{
+		Path:     dir,
+		Interval: testInterval,
+		OnChange: func() { atomic.AddInt64(&n, 1) },
+	})
+	defer stop()
+	time.Sleep(testInterval * 2)
+
+	require.NoError(t, os.Remove(filepath.Join(dir, "b.yaml")))
+	require.True(t, waitForCount(t, &n, 1, time.Second), "callback should fire when a file is removed")
+}
+
+func TestDirWatcher_DetectsModTimeChange(t *testing.T) {
+	dir := t.TempDir()
+	writeYAMLInDir(t, dir, "a.yaml", "a")
+
+	base := time.Now().Add(-1 * time.Hour).Truncate(time.Second)
+	require.NoError(t, os.Chtimes(filepath.Join(dir, "a.yaml"), base, base))
+
+	var n int64
+	stop := startDirWatcher(t, &DirWatcher{
+		Path:     dir,
+		Interval: testInterval,
+		OnChange: func() { atomic.AddInt64(&n, 1) },
+	})
+	defer stop()
+	time.Sleep(testInterval * 2)
+
+	require.NoError(t, os.Chtimes(filepath.Join(dir, "a.yaml"), base.Add(10*time.Second), base.Add(10*time.Second)))
+	require.True(t, waitForCount(t, &n, 1, time.Second), "callback should fire after mtime change")
+}
+
+func TestDirWatcher_IgnoresNonYAMLFiles(t *testing.T) {
+	dir := t.TempDir()
+	writeYAMLInDir(t, dir, "a.yaml", "a")
+
+	var n int64
+	stop := startDirWatcher(t, &DirWatcher{
+		Path:     dir,
+		Interval: testInterval,
+		OnChange: func() { atomic.AddInt64(&n, 1) },
+	})
+	defer stop()
+	time.Sleep(testInterval * 2)
+
+	// Adding a .txt file must not fire.
+	require.NoError(t, os.WriteFile(filepath.Join(dir, "notes.txt"), []byte("hi"), 0o644))
+	time.Sleep(testInterval * 4)
+	require.Equal(t, int64(0), atomic.LoadInt64(&n), "non-YAML files must be ignored")
+
+	// Adding a .yml file must fire.
+	writeYAMLInDir(t, dir, "b.yml", "b")
+	require.True(t, waitForCount(t, &n, 1, time.Second), "callback should fire for *.yml files")
+}
+
+func TestDirWatcher_MissingDirRecovers(t *testing.T) {
+	dir := t.TempDir()
+	writeYAMLInDir(t, dir, "a.yaml", "a")
+
+	var n int64
+	stop := startDirWatcher(t, &DirWatcher{
+		Path:     dir,
+		Interval: testInterval,
+		OnChange: func() { atomic.AddInt64(&n, 1) },
+	})
+	defer stop()
+	time.Sleep(testInterval * 2)
+
+	// Remove the directory. No fire expected on disappearance alone.
+	require.NoError(t, os.RemoveAll(dir))
+	time.Sleep(testInterval * 3)
+	require.Equal(t, int64(0), atomic.LoadInt64(&n), "directory removal alone must not fire")
+
+	// Recreate the directory and a YAML file; the recovery should fire.
+	require.NoError(t, os.MkdirAll(dir, 0o755))
+	writeYAMLInDir(t, dir, "recovered.yaml", "r")
+	require.True(t, waitForCount(t, &n, 1, time.Second), "callback should fire when dir returns with content")
+}
+
+func TestDirWatcher_EmptyDirSuppressedThenRecovers(t *testing.T) {
+	// Present-with-content → empty (all YAML removed, dir still exists)
+	// must stay quiet — treated as transient per the documented policy.
+	// The transition back to content fires.
+	dir := t.TempDir()
+	writeYAMLInDir(t, dir, "a.yaml", "a")
+
+	var n int64
+	stop := startDirWatcher(t, &DirWatcher{
+		Path:     dir,
+		Interval: testInterval,
+		OnChange: func() { atomic.AddInt64(&n, 1) },
+	})
+	defer stop()
+	time.Sleep(testInterval * 2)
+
+	// Remove the only YAML file. Dir still exists but is empty of YAML.
+	require.NoError(t, os.Remove(filepath.Join(dir, "a.yaml")))
+	time.Sleep(testInterval * 4)
+	require.Equal(t, int64(0), atomic.LoadInt64(&n), "emptying the directory must not fire")
+
+	// Add a YAML file back; transition to present-with-content fires.
+	writeYAMLInDir(t, dir, "c.yaml", "c")
+	require.True(t, waitForCount(t, &n, 1, time.Second), "callback should fire when content returns")
+}
+
+func TestDirWatcher_ContextCancelStopsRun(t *testing.T) {
+	dir := t.TempDir()
+	writeYAMLInDir(t, dir, "a.yaml", "a")
+
+	w := &DirWatcher{Path: dir, Interval: testInterval}
+	ctx, cancel := context.WithCancel(context.Background())
+	done := make(chan struct{})
+	go func() { w.Run(ctx); close(done) }()
+
+	time.Sleep(testInterval * 2)
+	cancel()
+	select {
+	case <-done:
+	case <-time.After(2 * time.Second):
+		t.Fatal("Run did not return within 2s of cancel")
+	}
+}

llama-swap.go

@@ -6,6 +6,7 @@ import (
 	"flag"
 	"fmt"
 	"log/slog"
+	"net"
 	"net/http"
 	"os"
 	"os/signal"
@@ -19,6 +20,7 @@ import (
 	"github.com/mostlygeek/llama-swap/internal/event"
 	"github.com/mostlygeek/llama-swap/internal/logmon"
 	"github.com/mostlygeek/llama-swap/internal/perf"
+	"github.com/mostlygeek/llama-swap/internal/process"
 	"github.com/mostlygeek/llama-swap/internal/server"
 	"github.com/mostlygeek/llama-swap/internal/shared"
 	"github.com/mostlygeek/llama-swap/internal/watcher"
@@ -53,7 +55,8 @@ var logTimeFormats = map[string]string{
 }
 
 func main() {
-	flagConfig := flag.String("config", "", "path to config file (required)")
+	flagConfig := flag.String("config", "", "path to config file")
+	flagConfigDir := flag.String("config-dir", "", "directory of *.yml/*.yaml config files (additive to -config)")
 	flagListen := flag.String("listen", "", "listen address (default :8080 or :8443 for TLS)")
 	flagCertFile := flag.String("tls-cert-file", "", "TLS certificate file")
 	flagKeyFile := flag.String("tls-key-file", "", "TLS key file")
@@ -66,8 +69,8 @@ func main() {
 		os.Exit(0)
 	}
 
-	if *flagConfig == "" {
-		slog.Error("-config is required")
+	if *flagConfig == "" && *flagConfigDir == "" {
+		slog.Error("at least one of -config or -config-dir must be provided")
 		os.Exit(1)
 	}
 
@@ -86,10 +89,9 @@ func main() {
 		}
 	}
 
-	configPath := *flagConfig
-	cfg, err := config.LoadConfig(configPath)
+	cfg, err := config.LoadConfigSources(*flagConfig, *flagConfigDir)
 	if err != nil {
-		slog.Error("failed to load config", "path", configPath, "error", err)
+		slog.Error("failed to load config", "config", *flagConfig, "config-dir", *flagConfigDir, "error", err)
 		os.Exit(1)
 	}
 
@@ -122,6 +124,13 @@ func main() {
 	applyLogSettings(cfg)
 	proxyLog.Debugf("PID: %d", os.Getpid())
 
+	// On Windows, bind the process tree to a Job Object so every upstream
+	// process is reaped when llama-swap exits — even on a forced kill. No-op
+	// elsewhere. Non-fatal: a failure just falls back to per-process teardown.
+	if err := process.SetupTreeCleanup(); err != nil {
+		proxyLog.Warnf("failed to set up process tree cleanup: %v", err)
+	}
+
 	// perfMon outlives config reloads; its config is updated in place.
 	var perfMon *perf.Monitor
 	if !cfg.Performance.Disabled {
@@ -178,7 +187,7 @@ func main() {
 
 		proxyLog.Info("reloading configuration")
 
-		newCfg, err := config.LoadConfig(configPath)
+		newCfg, err := config.LoadConfigSources(*flagConfig, *flagConfigDir)
 		if err != nil {
 			proxyLog.Warnf("failed to reload config: %v", err)
 			return
@@ -221,19 +230,37 @@ func main() {
 	defer watcherCancel()
 
 	if *flagWatchConfig {
-		absConfigPath, err := filepath.Abs(configPath)
-		if err != nil {
-			slog.Error("watch-config: failed to resolve config path", "error", err)
-			os.Exit(1)
-		}
 		proxyLog.Info("watching configuration for changes (poll-based, 2s interval)")
-		go func() {
-			(&configwatcher.Watcher{
-				Path:     absConfigPath,
-				Interval: configwatcher.DefaultInterval,
-				OnChange: reload,
-			}).Run(watcherCtx)
-		}()
+
+		if *flagConfig != "" {
+			absConfigPath, err := filepath.Abs(*flagConfig)
+			if err != nil {
+				slog.Error("watch-config: failed to resolve config path", "error", err)
+				os.Exit(1)
+			}
+			go func() {
+				(&configwatcher.Watcher{
+					Path:     absConfigPath,
+					Interval: configwatcher.DefaultInterval,
+					OnChange: reload,
+				}).Run(watcherCtx)
+			}()
+		}
+
+		if *flagConfigDir != "" {
+			absConfigDir, err := filepath.Abs(*flagConfigDir)
+			if err != nil {
+				slog.Error("watch-config: failed to resolve config-dir path", "error", err)
+				os.Exit(1)
+			}
+			go func() {
+				(&configwatcher.DirWatcher{
+					Path:     absConfigDir,
+					Interval: configwatcher.DefaultInterval,
+					OnChange: reload,
+				}).Run(watcherCtx)
+			}()
+		}
 	}
 
 	sigChan := make(chan os.Signal, 1)
@@ -254,6 +281,11 @@ func main() {
 		}
 	}()
 
+	if !shared.IsLoopbackAddr(listenAddr) {
+		_, port, _ := net.SplitHostPort(listenAddr)
+		proxyLog.Infof("llama-swap is reachable by all hosts on the network, use -listen localhost:%s to restrict to loopback only", port)
+	}
+
 	exitChan := make(chan struct{})
 
 	go func() {
@@ -267,6 +299,16 @@ func main() {
 				proxyLog.Infof("received signal %v, shutting down", sig)
 				watcherCancel()
 
+				// Backstop against a stalled shutdown: force the process to
+				// exit once the whole graceful sequence has had its full budget.
+				// On Windows the Job Object reaps upstream processes on exit, so
+				// a forced exit still cleans up rather than orphaning children.
+				go func() {
+					time.Sleep(shutdownTimeout + 5*time.Second)
+					proxyLog.Warnf("graceful shutdown exceeded %v, forcing exit", shutdownTimeout)
+					os.Exit(1)
+				}()
+
 				activeMu.RLock()
 				srv := activeSrv
 				activeMu.RUnlock()
@@ -275,13 +317,23 @@ func main() {
 				// drain without blocking on them for the full timeout.
 				srv.CloseStreams()
 
-				shutdownCtx, cancel := context.WithTimeout(context.Background(), shutdownTimeout)
+				// Both phases share a single deadline so total shutdown is
+				// bounded by shutdownTimeout rather than 2x it.
+				deadline := time.Now().Add(shutdownTimeout)
+				shutdownCtx, cancel := context.WithDeadline(context.Background(), deadline)
 				defer cancel()
 				if err := httpServer.Shutdown(shutdownCtx); err != nil {
 					proxyLog.Warnf("http server shutdown error: %v", err)
 				}
 
-				if err := srv.Shutdown(shutdownTimeout); err != nil {
+				// Clamp the remaining budget to a small positive value: a
+				// non-positive timeout makes the router fall back to its own
+				// healthCheckTimeout, which would defeat the shared deadline.
+				remaining := time.Until(deadline)
+				if remaining <= 0 {
+					remaining = time.Millisecond
+				}
+				if err := srv.Shutdown(remaining); err != nil {
 					proxyLog.Warnf("router shutdown error: %v", err)
 				}
 

proxy/.gitignore

@@ -1 +0,0 @@
-ui_dist/*

proxy/discardWriter.go

@@ -1,27 +0,0 @@
-package proxy
-
-import "net/http"
-
-// Custom discard writer that implements http.ResponseWriter but just discards everything
-type DiscardWriter struct {
-	header http.Header
-	status int
-}
-
-func (w *DiscardWriter) Header() http.Header {
-	if w.header == nil {
-		w.header = make(http.Header)
-	}
-	return w.header
-}
-
-func (w *DiscardWriter) Write(data []byte) (int, error) {
-	return len(data), nil
-}
-
-func (w *DiscardWriter) WriteHeader(code int) {
-	w.status = code
-}
-
-// Satisfy the http.Flusher interface for streaming responses
-func (w *DiscardWriter) Flush() {}

proxy/events.go

@@ -1,60 +0,0 @@
-package proxy
-
-// package level registry of the different event types
-
-const ProcessStateChangeEventID = 0x01
-const ChatCompletionStatsEventID = 0x02
-const ConfigFileChangedEventID = 0x03
-const ActivityLogEventID = 0x05
-const ModelPreloadedEventID = 0x06
-const InFlightRequestsEventID = 0x07
-
-type ProcessStateChangeEvent struct {
-	ProcessName string
-	NewState    ProcessState
-	OldState    ProcessState
-}
-
-func (e ProcessStateChangeEvent) Type() uint32 {
-	return ProcessStateChangeEventID
-}
-
-type ChatCompletionStats struct {
-	TokensGenerated int
-}
-
-func (e ChatCompletionStats) Type() uint32 {
-	return ChatCompletionStatsEventID
-}
-
-type ReloadingState int
-
-const (
-	ReloadingStateStart ReloadingState = iota
-	ReloadingStateEnd
-)
-
-type ConfigFileChangedEvent struct {
-	ReloadingState ReloadingState
-}
-
-func (e ConfigFileChangedEvent) Type() uint32 {
-	return ConfigFileChangedEventID
-}
-
-type ModelPreloadedEvent struct {
-	ModelName string
-	Success   bool
-}
-
-func (e ModelPreloadedEvent) Type() uint32 {
-	return ModelPreloadedEventID
-}
-
-type InFlightRequestsEvent struct {
-	Total int
-}
-
-func (e InFlightRequestsEvent) Type() uint32 {
-	return InFlightRequestsEventID
-}

proxy/helpers_test.go

@@ -1,304 +0,0 @@
-package proxy
-
-import (
-	"encoding/json"
-	"fmt"
-	"io"
-	"net/http"
-	"os"
-	"path/filepath"
-	"runtime"
-	"strconv"
-	"strings"
-	"sync"
-	"testing"
-	"time"
-
-	"github.com/gin-gonic/gin"
-	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/mostlygeek/llama-swap/internal/logmon"
-	"github.com/stretchr/testify/require"
-	"github.com/tidwall/gjson"
-	"gopkg.in/yaml.v3"
-)
-
-var (
-	nextTestPort        int = 12000
-	portMutex           sync.Mutex
-	testLogger          = logmon.NewWriter(os.Stdout)
-	simpleResponderPath = getSimpleResponderPath()
-)
-
-// Check if the binary exists
-func TestMain(m *testing.M) {
-	binaryPath := getSimpleResponderPath()
-	if _, err := os.Stat(binaryPath); os.IsNotExist(err) {
-		fmt.Printf("simple-responder not found at %s, did you `make simple-responder`?\n", binaryPath)
-		os.Exit(1)
-	}
-
-	gin.SetMode(gin.TestMode)
-
-	switch os.Getenv("LOG_LEVEL") {
-	case "debug":
-		testLogger.SetLogLevel(logmon.LevelDebug)
-	case "warn":
-		testLogger.SetLogLevel(logmon.LevelWarn)
-	case "info":
-		testLogger.SetLogLevel(logmon.LevelInfo)
-	default:
-		testLogger.SetLogLevel(logmon.LevelWarn)
-	}
-
-	m.Run()
-}
-
-// Helper function to get the binary path
-func getSimpleResponderPath() string {
-	goos := runtime.GOOS
-	goarch := runtime.GOARCH
-
-	if goos == "windows" {
-		return filepath.Join("..", "build", "simple-responder.exe")
-	} else {
-		return filepath.Join("..", "build", fmt.Sprintf("simple-responder_%s_%s", goos, goarch))
-	}
-}
-
-func getTestPort() int {
-	portMutex.Lock()
-	defer portMutex.Unlock()
-
-	port := nextTestPort
-	nextTestPort++
-
-	return port
-}
-
-// testConfigFromYAML substitutes {{RESPONDER}} with the simple-responder path and
-// loads through the real config pipeline (env vars, macros, port assignment, etc.)
-func testConfigFromYAML(t *testing.T, yamlTmpl string) config.Config {
-	t.Helper()
-	yamlStr := strings.ReplaceAll(yamlTmpl, "{{RESPONDER}}", filepath.ToSlash(simpleResponderPath))
-	cfg, err := config.LoadConfigFromReader(strings.NewReader(yamlStr))
-	require.NoError(t, err)
-	return cfg
-}
-
-func getTestSimpleResponderConfig(expectedMessage string) config.ModelConfig {
-	return getTestSimpleResponderConfigPort(expectedMessage, getTestPort())
-}
-
-func getTestSimpleResponderConfigPort(expectedMessage string, port int) config.ModelConfig {
-	// Convert path to forward slashes for cross-platform compatibility
-	// Windows handles forward slashes in paths correctly
-	cmdPath := filepath.ToSlash(simpleResponderPath)
-
-	// Create a YAML string with just the values we want to set
-	yamlStr := fmt.Sprintf(`
-cmd: '%s --port %d --silent --respond %s'
-proxy: "http://127.0.0.1:%d"
-`, cmdPath, port, expectedMessage, port)
-
-	var cfg config.ModelConfig
-	if err := yaml.Unmarshal([]byte(yamlStr), &cfg); err != nil {
-		panic(fmt.Sprintf("failed to unmarshal test config: %v in [%s]", err, yamlStr))
-	}
-
-	return cfg
-}
-
-// injectTestHandlers sets a testHandler on every Process in every ProcessGroup
-// of the given ProxyManager, bypassing subprocess launches. modelResponses maps
-// model IDs to their respond strings; if a model ID is not in the map, the model
-// ID itself is used.
-func injectTestHandlers(pm *ProxyManager, modelResponses map[string]string) {
-	for _, pg := range pm.processGroups {
-		for modelID, process := range pg.processes {
-			respond := modelID
-			if r, ok := modelResponses[modelID]; ok {
-				respond = r
-			}
-			process.testHandler = newTestHandler(respond)
-		}
-	}
-}
-
-// newTestHandler returns an http.Handler that mimics simple-responder's API.
-// It supports the endpoints that routing tests depend on, without launching
-// any subprocess or binding any port.
-func respondJSON(w http.ResponseWriter, respond string, bodyBytes []byte) {
-	w.Header().Set("Content-Type", "application/json")
-	json.NewEncoder(w).Encode(map[string]any{
-		"responseMessage":  respond,
-		"h_content_length": strconv.Itoa(len(bodyBytes)),
-		"request_body":     string(bodyBytes),
-		"usage": map[string]any{
-			"completion_tokens": 10, "prompt_tokens": 25, "total_tokens": 35,
-		},
-		"timings": map[string]any{
-			"prompt_n": 25, "prompt_ms": 13, "predicted_n": 10,
-			"predicted_ms": 17, "predicted_per_second": 10,
-		},
-	})
-}
-
-func newTestHandler(respond string) http.Handler {
-	mux := http.NewServeMux()
-
-	mux.HandleFunc("/v1/chat/completions", func(w http.ResponseWriter, r *http.Request) {
-		bodyBytes, _ := io.ReadAll(r.Body)
-		isStreaming := r.URL.Query().Get("stream") == "true"
-
-		if wait := r.URL.Query().Get("wait"); wait != "" {
-			if d, err := time.ParseDuration(wait); err == nil {
-				time.Sleep(d)
-			}
-		}
-
-		if isStreaming {
-			w.Header().Set("Content-Type", "text/event-stream")
-			w.Header().Set("Cache-Control", "no-cache")
-			w.Header().Set("Connection", "keep-alive")
-			flusher := w.(http.Flusher)
-
-			for i := 0; i < 10; i++ {
-				data, _ := json.Marshal(map[string]any{
-					"created": time.Now().Unix(),
-					"choices": []map[string]any{
-						{"index": 0, "delta": map[string]any{"content": "asdf"}, "finish_reason": nil},
-					},
-				})
-				fmt.Fprintf(w, "event: message\ndata: %s\n\n", data)
-				flusher.Flush()
-			}
-
-			finalData, _ := json.Marshal(map[string]any{
-				"usage": map[string]any{
-					"completion_tokens": 10, "prompt_tokens": 25, "total_tokens": 35,
-				},
-				"timings": map[string]any{
-					"prompt_n": 25, "prompt_ms": 13, "predicted_n": 10,
-					"predicted_ms": 17, "predicted_per_second": 10,
-				},
-			})
-			fmt.Fprintf(w, "event: message\ndata: %s\n\n", finalData)
-			flusher.Flush()
-
-			fmt.Fprintf(w, "event: message\ndata: [DONE]\n\n")
-			flusher.Flush()
-		} else {
-			respondJSON(w, respond, bodyBytes)
-		}
-	})
-
-	mux.HandleFunc("/v1/audio/speech", func(w http.ResponseWriter, r *http.Request) {
-		body, _ := io.ReadAll(r.Body)
-		modelName := gjson.GetBytes(body, "model").String()
-		if modelName != respond {
-			w.WriteHeader(http.StatusBadRequest)
-			json.NewEncoder(w).Encode(map[string]string{"error": fmt.Sprintf("Invalid model: %s, expected: %s", modelName, respond)})
-			return
-		}
-		json.NewEncoder(w).Encode(map[string]string{"message": "ok"})
-	})
-
-	mux.HandleFunc("/v1/completions", func(w http.ResponseWriter, r *http.Request) {
-		bodyBytes, _ := io.ReadAll(r.Body)
-		respondJSON(w, respond, bodyBytes)
-	})
-
-	for _, path := range []string{
-		"/chat/completions", "/completions",
-		"/responses", "/messages", "/messages/count_tokens",
-		"/embeddings", "/rerank", "/reranking",
-	} {
-		mux.HandleFunc(path, func(w http.ResponseWriter, r *http.Request) {
-			bodyBytes, _ := io.ReadAll(r.Body)
-			respondJSON(w, respond, bodyBytes)
-		})
-	}
-
-	mux.HandleFunc("/completion", func(w http.ResponseWriter, r *http.Request) {
-		w.Header().Set("Content-Type", "application/json")
-		json.NewEncoder(w).Encode(map[string]any{
-			"responseMessage": respond,
-			"usage": map[string]any{
-				"completion_tokens": 10, "prompt_tokens": 25, "total_tokens": 35,
-			},
-		})
-	})
-
-	mux.HandleFunc("/v1/audio/transcriptions", func(w http.ResponseWriter, r *http.Request) {
-		if err := r.ParseMultipartForm(10 << 20); err != nil {
-			w.WriteHeader(http.StatusBadRequest)
-			json.NewEncoder(w).Encode(map[string]string{"error": fmt.Sprintf("Error parsing multipart form: %s", err)})
-			return
-		}
-		model := r.FormValue("model")
-		if model == "" {
-			w.WriteHeader(http.StatusBadRequest)
-			json.NewEncoder(w).Encode(map[string]string{"error": "Missing model parameter"})
-			return
-		}
-		file, _, err := r.FormFile("file")
-		if err != nil {
-			w.WriteHeader(http.StatusBadRequest)
-			json.NewEncoder(w).Encode(map[string]string{"error": fmt.Sprintf("Error getting file: %s", err)})
-			return
-		}
-		fileBytes, _ := io.ReadAll(file)
-		file.Close()
-		json.NewEncoder(w).Encode(map[string]any{
-			"text":             fmt.Sprintf("The length of the file is %d bytes", len(fileBytes)),
-			"model":            model,
-			"h_content_type":   r.Header.Get("Content-Type"),
-			"h_content_length": r.Header.Get("Content-Length"),
-		})
-	})
-
-	mux.HandleFunc("/v1/audio/voices", func(w http.ResponseWriter, r *http.Request) {
-		model := r.URL.Query().Get("model")
-		json.NewEncoder(w).Encode(map[string]any{
-			"voices": []string{"voice1"}, "model": model,
-		})
-	})
-
-	mux.HandleFunc("/test", func(w http.ResponseWriter, r *http.Request) {
-		w.Header().Set("Content-Type", "text/plain")
-		fmt.Fprint(w, respond)
-	})
-
-	mux.HandleFunc("/", func(w http.ResponseWriter, r *http.Request) {
-		if r.URL.Path != "/" {
-			http.NotFound(w, r)
-			return
-		}
-		w.Header().Set("Content-Type", "text/plain")
-		fmt.Fprintf(w, "%s %s", r.Method, r.URL.Path)
-	})
-
-	mux.HandleFunc("/sdapi/v1/txt2img", func(w http.ResponseWriter, r *http.Request) {
-		body, _ := io.ReadAll(r.Body)
-		modelName := gjson.GetBytes(body, "model").String()
-		json.NewEncoder(w).Encode(map[string]any{
-			"model": modelName, "images": []string{},
-		})
-	})
-
-	mux.HandleFunc("/sdapi/v1/img2img", func(w http.ResponseWriter, r *http.Request) {
-		body, _ := io.ReadAll(r.Body)
-		modelName := gjson.GetBytes(body, "model").String()
-		json.NewEncoder(w).Encode(map[string]any{
-			"model": modelName, "images": []string{},
-		})
-	})
-
-	mux.HandleFunc("/sdapi/v1/loras", func(w http.ResponseWriter, r *http.Request) {
-		json.NewEncoder(w).Encode(map[string]any{
-			"loras": []string{},
-		})
-	})
-
-	return mux
-}

proxy/matrix.go

@@ -1,330 +0,0 @@
-package proxy
-
-import (
-	"fmt"
-	"net/http"
-	"slices"
-	"sort"
-	"sync"
-
-	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/mostlygeek/llama-swap/internal/logmon"
-)
-
-// MatrixSolver contains pure swap-decision logic with no Process dependencies.
-// It is safe for concurrent reads after construction.
-type MatrixSolver struct {
-	expandedSets []config.ExpandedSet // all valid model combinations
-	evictCosts   map[string]int       // real model name -> eviction cost (default 1)
-	modelToSets  map[string][]int     // model name -> indices into expandedSets
-}
-
-// NewMatrixSolver builds a solver from expanded sets and eviction costs.
-func NewMatrixSolver(expandedSets []config.ExpandedSet, evictCosts map[string]int) *MatrixSolver {
-	modelToSets := make(map[string][]int)
-	for i, es := range expandedSets {
-		for _, model := range es.Models {
-			modelToSets[model] = append(modelToSets[model], i)
-		}
-	}
-
-	return &MatrixSolver{
-		expandedSets: expandedSets,
-		evictCosts:   evictCosts,
-		modelToSets:  modelToSets,
-	}
-}
-
-// SolveResult describes what the solver decided.
-type SolveResult struct {
-	Evict     []string // running models that must be stopped
-	TargetSet []string // the chosen set of models (for informational purposes)
-	SetName   string   // name of the chosen set
-	DSL       string   // original DSL expression for the chosen set
-	TotalCost int      // total eviction cost
-}
-
-// Solve determines which models to evict when a model is requested.
-//
-// Algorithm:
-//  1. If requestedModel is already running, no eviction needed.
-//  2. Find all sets containing requestedModel.
-//  3. If no sets found, the model runs alone; evict all running models.
-//  4. For each candidate set, compute cost = sum of evict_costs for running
-//     models NOT in that set.
-//  5. Pick lowest cost. Ties broken by definition order (index in expandedSets).
-//  6. Return models to evict and the chosen set.
-func (s *MatrixSolver) Solve(requestedModel string, runningModels []string) (SolveResult, error) {
-	// If already running, nothing to do (but fill in set info for logging)
-	if slices.Contains(runningModels, requestedModel) {
-		setName, dsl := s.findMatchingSet(requestedModel, runningModels)
-		return SolveResult{
-			TargetSet: runningModels,
-			SetName:   setName,
-			DSL:       dsl,
-		}, nil
-	}
-
-	candidateIndices := s.modelToSets[requestedModel]
-
-	// Model not in any set: runs alone, evict everything
-	if len(candidateIndices) == 0 {
-		evict := make([]string, len(runningModels))
-		copy(evict, runningModels)
-		return SolveResult{
-			Evict:     evict,
-			TargetSet: []string{requestedModel},
-		}, nil
-	}
-
-	// Find the cheapest candidate set
-	bestCost := -1
-	bestIdx := -1
-
-	for _, idx := range candidateIndices {
-		setModels := s.expandedSets[idx].Models
-		cost := 0
-		for _, running := range runningModels {
-			if !slices.Contains(setModels, running) {
-				cost += s.evictCost(running)
-			}
-		}
-
-		if bestCost < 0 || cost < bestCost || (cost == bestCost && idx < bestIdx) {
-			bestCost = cost
-			bestIdx = idx
-		}
-	}
-
-	// Determine which running models to evict
-	chosen := s.expandedSets[bestIdx]
-	var evict []string
-	for _, running := range runningModels {
-		if !slices.Contains(chosen.Models, running) {
-			evict = append(evict, running)
-		}
-	}
-
-	return SolveResult{
-		Evict:     evict,
-		TargetSet: chosen.Models,
-		SetName:   chosen.SetName,
-		DSL:       chosen.DSL,
-		TotalCost: bestCost,
-	}, nil
-}
-
-// findMatchingSet finds the expanded set that contains all running models.
-// Returns the set name and DSL, or empty strings if no match.
-func (s *MatrixSolver) findMatchingSet(requestedModel string, runningModels []string) (string, string) {
-	for _, idx := range s.modelToSets[requestedModel] {
-		set := s.expandedSets[idx]
-		allInSet := true
-		for _, m := range runningModels {
-			if !slices.Contains(set.Models, m) {
-				allInSet = false
-				break
-			}
-		}
-		if allInSet {
-			return set.SetName, set.DSL
-		}
-	}
-	return "", ""
-}
-
-func (s *MatrixSolver) evictCost(model string) int {
-	if cost, ok := s.evictCosts[model]; ok {
-		return cost
-	}
-	return 1
-}
-
-// Matrix manages processes using solver-based swap logic.
-type Matrix struct {
-	sync.Mutex
-	solver         *MatrixSolver
-	processes      map[string]*Process // all processes keyed by real model name
-	config         config.Config
-	proxyLogger    *logmon.Monitor
-	upstreamLogger *logmon.Monitor
-
-	// inflight tracks ProxyRequest calls that have released m.Lock but may
-	// not yet have incremented Process.inFlightRequests. A concurrent
-	// request that needs to evict models waits for inflight to drain under
-	// m.Lock before stopping anything. Without this, a request that
-	// released m.Lock but has not yet reached Process.inFlightRequests.Add(1)
-	// races with Stop()'s Wait() and can be killed mid-request.
-	inflight sync.WaitGroup
-
-	// testDelayFastPath is a test-only hook invoked in the no-eviction path
-	// after m.Lock is released but before the request is dispatched to
-	// Process.ProxyRequest. Tests use it to park a request at the exact
-	// race window to deterministically reproduce the race.
-	testDelayFastPath func()
-}
-
-// NewMatrix creates a Matrix from config. It creates a Process for every
-// model defined in the config (any model can run alone even if not in a set).
-func NewMatrix(cfg config.Config, proxyLogger, upstreamLogger *logmon.Monitor) *Matrix {
-	processes := make(map[string]*Process)
-	for modelID, modelConfig := range cfg.Models {
-		processLogger := logmon.NewWriter(upstreamLogger)
-		process := NewProcess(modelID, cfg.HealthCheckTimeout, modelConfig, processLogger, proxyLogger)
-		processes[modelID] = process
-	}
-
-	evictCosts := cfg.Matrix.ResolvedEvictCosts()
-
-	return &Matrix{
-		solver:         NewMatrixSolver(cfg.ExpandedSets, evictCosts),
-		processes:      processes,
-		config:         cfg,
-		proxyLogger:    proxyLogger,
-		upstreamLogger: upstreamLogger,
-	}
-}
-
-// ProxyRequest handles the swap logic and proxies the request to the model.
-func (m *Matrix) ProxyRequest(modelID string, w http.ResponseWriter, r *http.Request) error {
-	process, ok := m.processes[modelID]
-	if !ok {
-		return fmt.Errorf("model %s not found in matrix", modelID)
-	}
-
-	m.Lock()
-	running := m.runningModels()
-	result, err := m.solver.Solve(modelID, running)
-	if err != nil {
-		m.Unlock()
-		return fmt.Errorf("matrix solver error: %w", err)
-	}
-
-	// Log solver decision
-	if len(result.Evict) > 0 {
-		m.proxyLogger.Infof("Matrix: model=%s set=%s dsl=%q evict=%v target=%v cost=%d",
-			modelID, result.SetName, result.DSL, result.Evict, result.TargetSet, result.TotalCost)
-	} else if len(running) == 0 {
-		m.proxyLogger.Infof("Matrix: model=%s starting (no models running)", modelID)
-	} else {
-		m.proxyLogger.Debugf("Matrix: model=%s already running in set=%s dsl=%q", modelID, result.SetName, result.DSL)
-	}
-
-	// Evict models that need to be stopped
-	if len(result.Evict) > 0 {
-		// Wait for any in-flight ProxyRequest calls to register on their
-		// Process before stopping anything. Without this, a request that
-		// released m.Lock but has not yet incremented
-		// Process.inFlightRequests races with Stop() and can be killed
-		// mid-request.
-		m.inflight.Wait()
-
-		var wg sync.WaitGroup
-		for _, evictModel := range result.Evict {
-			if p, exists := m.processes[evictModel]; exists {
-				wg.Add(1)
-				go func(p *Process) {
-					defer wg.Done()
-					p.Stop()
-				}(p)
-			}
-		}
-		wg.Wait()
-	}
-
-	// Register this request in inflight before releasing m.Lock so a
-	// concurrent eviction will wait for it to complete.
-	m.inflight.Add(1)
-	defer m.inflight.Done()
-	isFastPath := len(result.Evict) == 0
-	m.Unlock()
-
-	if isFastPath && m.testDelayFastPath != nil {
-		m.testDelayFastPath()
-	}
-
-	// Proxy the request (Process handles on-demand start)
-	process.ProxyRequest(w, r)
-	return nil
-}
-
-// StopProcesses stops all running processes.
-func (m *Matrix) StopProcesses(strategy StopStrategy) {
-	m.Lock()
-	defer m.Unlock()
-
-	var wg sync.WaitGroup
-	for _, process := range m.processes {
-		wg.Add(1)
-		go func(p *Process) {
-			defer wg.Done()
-			switch strategy {
-			case StopImmediately:
-				p.StopImmediately()
-			default:
-				p.Stop()
-			}
-		}(process)
-	}
-	wg.Wait()
-}
-
-// StopProcess stops a single process by model ID.
-func (m *Matrix) StopProcess(modelID string, strategy StopStrategy) error {
-	process, ok := m.processes[modelID]
-	if !ok {
-		return fmt.Errorf("process not found for %s", modelID)
-	}
-
-	switch strategy {
-	case StopImmediately:
-		process.StopImmediately()
-	default:
-		process.Stop()
-	}
-	return nil
-}
-
-// Shutdown shuts down all processes.
-func (m *Matrix) Shutdown() {
-	var wg sync.WaitGroup
-	for _, process := range m.processes {
-		wg.Add(1)
-		go func(p *Process) {
-			defer wg.Done()
-			p.Shutdown()
-		}(process)
-	}
-	wg.Wait()
-}
-
-// RunningModels returns model names currently in an active (non-stopped) state.
-func (m *Matrix) RunningModels() []string {
-	m.Lock()
-	defer m.Unlock()
-	return m.runningModels()
-}
-
-// runningModels returns running model names (caller must hold lock).
-func (m *Matrix) runningModels() []string {
-	var running []string
-	for id, process := range m.processes {
-		if process.CurrentState() != StateStopped && process.CurrentState() != StateShutdown {
-			running = append(running, id)
-		}
-	}
-	sort.Strings(running)
-	return running
-}
-
-// GetProcess returns the Process for a model.
-func (m *Matrix) GetProcess(modelID string) (*Process, bool) {
-	p, ok := m.processes[modelID]
-	return p, ok
-}
-
-// HasModel returns true if the model is managed by this matrix.
-func (m *Matrix) HasModel(modelID string) bool {
-	_, ok := m.processes[modelID]
-	return ok
-}

proxy/matrix_test.go

@@ -1,349 +0,0 @@
-package proxy
-
-import (
-	"net/http"
-	"net/http/httptest"
-	"runtime"
-	"testing"
-	"time"
-
-	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-// Helper to build expanded sets for solver tests
-func makeExpandedSets(sets ...struct {
-	name   string
-	models []string
-}) []config.ExpandedSet {
-	var result []config.ExpandedSet
-	for _, s := range sets {
-		result = append(result, config.ExpandedSet{
-			SetName: s.name,
-			Models:  s.models,
-		})
-	}
-	return result
-}
-
-func es(name string, models ...string) struct {
-	name   string
-	models []string
-} {
-	return struct {
-		name   string
-		models []string
-	}{name, models}
-}
-
-func TestMatrixSolver_AlreadyRunning(t *testing.T) {
-	solver := NewMatrixSolver(
-		makeExpandedSets(es("s1", "a", "b")),
-		nil,
-	)
-
-	result, err := solver.Solve("a", []string{"a"})
-	require.NoError(t, err)
-	assert.Empty(t, result.Evict)
-	assert.Equal(t, []string{"a"}, result.TargetSet)
-	assert.Equal(t, "s1", result.SetName)
-}
-
-func TestMatrixSolver_NotInAnySet_RunsAlone(t *testing.T) {
-	solver := NewMatrixSolver(
-		makeExpandedSets(es("s1", "a", "b")),
-		nil,
-	)
-
-	// Model "c" not in any set
-	result, err := solver.Solve("c", []string{"a", "b"})
-	require.NoError(t, err)
-	assert.ElementsMatch(t, []string{"a", "b"}, result.Evict)
-	assert.Equal(t, []string{"c"}, result.TargetSet)
-}
-
-func TestMatrixSolver_NotInAnySet_NothingRunning(t *testing.T) {
-	solver := NewMatrixSolver(
-		makeExpandedSets(es("s1", "a", "b")),
-		nil,
-	)
-
-	result, err := solver.Solve("c", []string{})
-	require.NoError(t, err)
-	assert.Empty(t, result.Evict)
-	assert.Equal(t, []string{"c"}, result.TargetSet)
-}
-
-func TestMatrixSolver_SingleSet_EvictsNonMembers(t *testing.T) {
-	// Set: [a, b]. Request a when b and c are running.
-	solver := NewMatrixSolver(
-		makeExpandedSets(es("s1", "a", "b")),
-		nil,
-	)
-
-	result, err := solver.Solve("a", []string{"b", "c"})
-	require.NoError(t, err)
-	// c is not in the set, so it gets evicted. b is in the set, so it stays.
-	assert.Equal(t, []string{"c"}, result.Evict)
-	assert.Equal(t, []string{"a", "b"}, result.TargetSet)
-}
-
-func TestMatrixSolver_PicksLowestCost(t *testing.T) {
-	// Two sets containing model "a":
-	// s1: [a, v]  — if v is running, cost=0; if L is running, cost=30
-	// s2: [a, L]  — if L is running, cost=0; if v is running, cost=50
-	solver := NewMatrixSolver(
-		makeExpandedSets(
-			es("s1", "a", "v"),
-			es("s2", "a", "L"),
-		),
-		map[string]int{"v": 50, "L": 30},
-	)
-
-	// v is running. Switching to a:
-	// s1 cost: v is in s1, so 0
-	// s2 cost: v is NOT in s2, so 50
-	// => pick s1
-	result, err := solver.Solve("a", []string{"v"})
-	require.NoError(t, err)
-	assert.Empty(t, result.Evict)
-	assert.Equal(t, []string{"a", "v"}, result.TargetSet)
-
-	// L is running. Switching to a:
-	// s1 cost: L is NOT in s1, so 30
-	// s2 cost: L is in s2, so 0
-	// => pick s2
-	result, err = solver.Solve("a", []string{"L"})
-	require.NoError(t, err)
-	assert.Empty(t, result.Evict)
-	assert.Equal(t, []string{"a", "L"}, result.TargetSet)
-}
-
-func TestMatrixSolver_TieBreakingByDefinitionOrder(t *testing.T) {
-	// Two sets with identical cost. Definition order should win.
-	solver := NewMatrixSolver(
-		makeExpandedSets(
-			es("s1", "a", "x"),
-			es("s2", "a", "y"),
-		),
-		nil,
-	)
-
-	// Nothing running, both sets cost 0. s1 is first.
-	result, err := solver.Solve("a", []string{})
-	require.NoError(t, err)
-	assert.Empty(t, result.Evict)
-	assert.Equal(t, []string{"a", "x"}, result.TargetSet)
-}
-
-func TestMatrixSolver_EvictCostPreservesExpensive(t *testing.T) {
-	// Model "v" costs 50 to evict, "m" costs 1 (default).
-	// Sets: [g,v], [g,m]
-	// Running: v, m. Request g.
-	// s1=[g,v]: evict m (cost 1), keep v
-	// s2=[g,m]: evict v (cost 50), keep m
-	// => pick s1
-	solver := NewMatrixSolver(
-		makeExpandedSets(
-			es("s1", "g", "v"),
-			es("s2", "g", "m"),
-		),
-		map[string]int{"v": 50},
-	)
-
-	result, err := solver.Solve("g", []string{"v", "m"})
-	require.NoError(t, err)
-	assert.Equal(t, []string{"m"}, result.Evict)
-	assert.Equal(t, []string{"g", "v"}, result.TargetSet)
-}
-
-func TestMatrixSolver_NothingRunning(t *testing.T) {
-	solver := NewMatrixSolver(
-		makeExpandedSets(
-			es("s1", "g", "v"),
-			es("s2", "q", "v"),
-		),
-		nil,
-	)
-
-	result, err := solver.Solve("g", []string{})
-	require.NoError(t, err)
-	assert.Empty(t, result.Evict)
-	assert.Equal(t, []string{"g", "v"}, result.TargetSet)
-}
-
-// TestMatrix_ProxyRequestSwapRaceAgainstFastPath verifies that an eviction
-// cannot stop a process while an in-flight ProxyRequest for that process is
-// still in the [m.Unlock, Process.inFlightRequests.Add(1)] window. Without
-// matrix-level inflight tracking, the eviction's Stop() races with the
-// pending request and kills it mid-start.
-func TestMatrix_ProxyRequestSwapRaceAgainstFastPath(t *testing.T) {
-	cfg := config.Config{
-		HealthCheckTimeout: 15,
-		Models: map[string]config.ModelConfig{
-			"model1": getTestSimpleResponderConfig("model1"),
-			"model2": getTestSimpleResponderConfig("model2"),
-		},
-		ExpandedSets: []config.ExpandedSet{
-			{SetName: "s1", Models: []string{"model1"}},
-			{SetName: "s2", Models: []string{"model2"}},
-		},
-		Matrix: &config.MatrixConfig{},
-	}
-
-	m := NewMatrix(cfg, testLogger, testLogger)
-	defer m.StopProcesses(StopImmediately)
-
-	// Bypass real subprocesses so the test is fast and deterministic.
-	m.processes["model1"].testHandler = newTestHandler("model1")
-	m.processes["model2"].testHandler = newTestHandler("model2")
-
-	// Prime: run a request through model1 so it reaches StateReady and
-	// subsequent requests take the no-eviction path.
-	primeReq := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-	primeW := httptest.NewRecorder()
-	require.NoError(t, m.ProxyRequest("model1", primeW, primeReq))
-	require.Equal(t, http.StatusOK, primeW.Code)
-	require.Equal(t, StateReady, m.processes["model1"].CurrentState())
-	require.Equal(t, StateStopped, m.processes["model2"].CurrentState())
-
-	// Install fast-path hook that signals arrival and waits for release.
-	// This parks R2 at the race window — after m.Lock is released but
-	// before Process.inFlightRequests.Add(1).
-	r2Reached := make(chan struct{})
-	r2Release := make(chan struct{})
-	m.testDelayFastPath = func() {
-		close(r2Reached)
-		<-r2Release
-	}
-
-	// R2: no-eviction request for model1. Will pause at the hook.
-	r2Done := make(chan struct{})
-	w2 := httptest.NewRecorder()
-	go func() {
-		defer close(r2Done)
-		req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-		assert.NoError(t, m.ProxyRequest("model1", w2, req))
-	}()
-
-	// Deterministically wait for R2 to reach the race window.
-	<-r2Reached
-
-	// R3: request for model2 which requires evicting model1. Must wait for
-	// R2 to finish before touching model1.
-	r3Done := make(chan struct{})
-	w3 := httptest.NewRecorder()
-	go func() {
-		defer close(r3Done)
-		req := httptest.NewRequest("POST", "/v1/chat/completions", nil)
-		assert.NoError(t, m.ProxyRequest("model2", w3, req))
-	}()
-
-	// Spin until R3 has acquired m.Lock and entered the eviction path. In
-	// the fixed code, R3 then blocks on m.inflight.Wait() while still
-	// holding the lock, so TryLock keeps failing.
-	for m.TryLock() {
-		m.Unlock()
-		runtime.Gosched()
-	}
-
-	// Bounded poll: give R3 a chance to demonstrate the bug by mutating
-	// state. In the fixed code R3 is blocked and nothing changes; in the
-	// buggy code R3 will Stop() model1 and start model2 within microseconds.
-	deadline := time.Now().Add(100 * time.Millisecond)
-	for time.Now().Before(deadline) {
-		if m.processes["model1"].CurrentState() != StateReady ||
-			m.processes["model2"].CurrentState() != StateStopped {
-			break
-		}
-		done := false
-		select {
-		case <-r3Done:
-			done = true
-		default:
-		}
-		if done {
-			break
-		}
-		runtime.Gosched()
-	}
-
-	// Invariant: R3 must be blocked while R2 is still in flight.
-	select {
-	case <-r3Done:
-		t.Fatal("eviction completed while in-flight request was still pending — race not prevented")
-	default:
-	}
-	assert.Equal(t, StateReady, m.processes["model1"].CurrentState(),
-		"model1 must stay Ready while an in-flight request is pending")
-	assert.Equal(t, StateStopped, m.processes["model2"].CurrentState(),
-		"model2 must not be started until R2 finishes and model1 is evicted")
-
-	// Release R2 and let both requests finish.
-	close(r2Release)
-	<-r2Done
-	<-r3Done
-
-	assert.Equal(t, http.StatusOK, w2.Code)
-	assert.Contains(t, w2.Body.String(), "model1")
-	assert.Equal(t, http.StatusOK, w3.Code)
-	assert.Contains(t, w3.Body.String(), "model2")
-}
-
-func TestMatrixSolver_FullScenario(t *testing.T) {
-	// Simulates the example config:
-	// standard: [g,v], [q,v], [m,v]
-	// with_rerank: [g,v,e], [q,v,e]
-	// creative: [g,sd], [q,sd]
-	// full: [L]
-	solver := NewMatrixSolver(
-		makeExpandedSets(
-			es("standard", "g", "v"),
-			es("standard", "q", "v"),
-			es("standard", "m", "v"),
-			es("with_rerank", "e", "g", "v"),
-			es("with_rerank", "e", "q", "v"),
-			es("creative", "g", "sd"),
-			es("creative", "q", "sd"),
-			es("full", "L"),
-		),
-		map[string]int{"v": 50, "L": 30, "whisper": 10},
-	)
-
-	// Running: g, v. Request q.
-	// standard[q,v]: evict g (cost 1), keep v. Total: 1.
-	// with_rerank[q,v,e]: evict g (cost 1), keep v. Total: 1.
-	// => tie, pick first by definition order = standard[q,v]
-	result, err := solver.Solve("q", []string{"g", "v"})
-	require.NoError(t, err)
-	assert.Equal(t, []string{"g"}, result.Evict)
-	assert.Equal(t, []string{"q", "v"}, result.TargetSet)
-
-	// Running: g, v. Request L.
-	// full[L]: evict g (cost 1) + v (cost 50). Total: 51.
-	// Only one set contains L, so pick it.
-	result, err = solver.Solve("L", []string{"g", "v"})
-	require.NoError(t, err)
-	assert.ElementsMatch(t, []string{"g", "v"}, result.Evict)
-	assert.Equal(t, []string{"L"}, result.TargetSet)
-
-	// Running: g, v. Request sd.
-	// creative[g,sd]: evict v (cost 50). Total: 50.
-	// creative[q,sd]: evict g (cost 1) + v (cost 50). Total: 51.
-	// => pick creative[g,sd]
-	result, err = solver.Solve("sd", []string{"g", "v"})
-	require.NoError(t, err)
-	assert.Equal(t, []string{"v"}, result.Evict)
-	assert.Equal(t, []string{"g", "sd"}, result.TargetSet)
-
-	// Running: q, v, e. Request g.
-	// standard[g,v]: evict q (1) + e (1). Total: 2.
-	// with_rerank[g,v,e]: evict q (1). Total: 1.
-	// creative[g,sd]: evict q (1) + v (50) + e (1). Total: 52.
-	// => pick with_rerank[g,v,e]
-	result, err = solver.Solve("g", []string{"e", "q", "v"})
-	require.NoError(t, err)
-	assert.Equal(t, []string{"q"}, result.Evict)
-	assert.Equal(t, []string{"e", "g", "v"}, result.TargetSet)
-}

proxy/metrics_monitor.go

@@ -1,689 +0,0 @@
-package proxy
-
-import (
-	"bytes"
-	"compress/flate"
-	"compress/gzip"
-	"encoding/json"
-	"fmt"
-	"io"
-	"net/http"
-	"strings"
-	"sync"
-	"time"
-
-	"github.com/fxamacker/cbor/v2"
-	"github.com/gin-gonic/gin"
-	"github.com/klauspost/compress/zstd"
-	"github.com/mostlygeek/llama-swap/internal/cache"
-	"github.com/mostlygeek/llama-swap/internal/event"
-	"github.com/mostlygeek/llama-swap/internal/logmon"
-	"github.com/mostlygeek/llama-swap/internal/ring"
-	"github.com/tidwall/gjson"
-)
-
-// zstdEncOptions are the shared zstd encoder options for maximum compression.
-var zstdEncOptions = []zstd.EOption{
-	zstd.WithEncoderLevel(zstd.SpeedBetterCompression),
-}
-
-// zstdDecOptions are the shared zstd decoder options.
-var zstdDecOptions = []zstd.DOption{}
-
-// zstdEncPool pools zstd.Encoder instances to reduce allocations.
-var zstdEncPool = &sync.Pool{
-	New: func() interface{} {
-		enc, _ := zstd.NewWriter(nil, zstdEncOptions...)
-		return enc
-	},
-}
-
-// zstdDecPool pools zstd.Decoder instances to reduce allocations.
-var zstdDecPool = &sync.Pool{
-	New: func() interface{} {
-		dec, _ := zstd.NewReader(nil, zstdDecOptions...)
-		return dec
-	},
-}
-
-// compressCapture marshals a ReqRespCapture to CBOR and compresses it with zstd.
-// Returns compressed bytes and the original CBOR byte count for logging.
-func compressCapture(c *ReqRespCapture) ([]byte, int, error) {
-	cborBytes, err := cbor.Marshal(c)
-	if err != nil {
-		return nil, 0, fmt.Errorf("marshal capture: %w", err)
-	}
-	zenc := zstdEncPool.Get().(*zstd.Encoder)
-	defer zstdEncPool.Put(zenc)
-	return zenc.EncodeAll(cborBytes, nil), len(cborBytes), nil
-}
-
-// decompressCapture decompresses zstd-compressed CBOR and unmarshals it into a ReqRespCapture.
-func decompressCapture(data []byte) (*ReqRespCapture, error) {
-	dec := zstdDecPool.Get().(*zstd.Decoder)
-	defer zstdDecPool.Put(dec)
-	cborBytes, err := dec.DecodeAll(data, nil)
-	if err != nil {
-		return nil, fmt.Errorf("decompress capture: %w", err)
-	}
-	var capture ReqRespCapture
-	if err := cbor.Unmarshal(cborBytes, &capture); err != nil {
-		return nil, fmt.Errorf("unmarshal capture: %w", err)
-	}
-	return &capture, nil
-}
-
-// TokenMetrics holds token usage and performance metrics
-type TokenMetrics struct {
-	CachedTokens    int     `json:"cache_tokens"`
-	InputTokens     int     `json:"input_tokens"`
-	OutputTokens    int     `json:"output_tokens"`
-	PromptPerSecond float64 `json:"prompt_per_second"`
-	TokensPerSecond float64 `json:"tokens_per_second"`
-}
-
-// ActivityLogEntry represents parsed token statistics from llama-server logs
-type ActivityLogEntry struct {
-	ID              int          `json:"id"`
-	Timestamp       time.Time    `json:"timestamp"`
-	Model           string       `json:"model"`
-	ReqPath         string       `json:"req_path"`
-	RespContentType string       `json:"resp_content_type"`
-	RespStatusCode  int          `json:"resp_status_code"`
-	Tokens          TokenMetrics `json:"tokens"`
-	DurationMs      int          `json:"duration_ms"`
-	HasCapture      bool         `json:"has_capture"`
-}
-
-type ReqRespCapture struct {
-	ID          int               `json:"id"`
-	ReqPath     string            `json:"req_path"`
-	ReqHeaders  map[string]string `json:"req_headers"`
-	ReqBody     []byte            `json:"req_body"`
-	RespHeaders map[string]string `json:"resp_headers"`
-	RespBody    []byte            `json:"resp_body"`
-}
-
-// ActivityLogEvent represents a token metrics event
-type ActivityLogEvent struct {
-	Metrics ActivityLogEntry
-}
-
-func (e ActivityLogEvent) Type() uint32 {
-	return ActivityLogEventID // defined in events.go
-}
-
-// metricsMonitor parses llama-server output for token statistics
-type metricsMonitor struct {
-	mu      sync.RWMutex
-	metrics ring.Buffer[ActivityLogEntry]
-	nextID  int
-	logger  *logmon.Monitor
-
-	// capture fields
-	enableCaptures bool
-	captureCache   *cache.Cache // zstd-compressed CBOR of ReqRespCapture
-}
-
-// newMetricsMonitor creates a new metricsMonitor. captureBufferMB is the
-// capture buffer size in megabytes; 0 disables captures.
-func newMetricsMonitor(logger *logmon.Monitor, maxMetrics int, captureBufferMB int) *metricsMonitor {
-	mm := &metricsMonitor{
-		logger:         logger,
-		metrics:        ring.NewBuffer[ActivityLogEntry](maxMetrics),
-		enableCaptures: captureBufferMB > 0,
-	}
-	if captureBufferMB > 0 {
-		mm.captureCache = cache.New(captureBufferMB * 1024 * 1024)
-	}
-	return mm
-}
-
-// queueMetrics adds a new metric to the collection without emitting an event.
-// Returns the assigned metric ID. Call emitMetric after capture setup.
-func (mp *metricsMonitor) queueMetrics(metric ActivityLogEntry) int {
-	mp.mu.Lock()
-	defer mp.mu.Unlock()
-
-	metric.ID = mp.nextID
-	mp.nextID++
-	mp.metrics.Push(metric)
-	return metric.ID
-}
-
-// emitMetric publishes an ActivityLogEvent for the given metric.
-func (mp *metricsMonitor) emitMetric(metric ActivityLogEntry) {
-	event.Emit(ActivityLogEvent{Metrics: metric})
-}
-
-// addCapture compresses and stores a capture in the cache.
-// Returns true if the capture was stored, false otherwise.
-func (mp *metricsMonitor) addCapture(capture ReqRespCapture) bool {
-	if !mp.enableCaptures {
-		return false
-	}
-
-	compressed, uncompressedBytes, err := compressCapture(&capture)
-	if err != nil {
-		mp.logger.Warnf("failed to compress capture: %v, skipping", err)
-		return false
-	}
-
-	if err := mp.captureCache.Add(capture.ID, compressed); err != nil {
-		mp.logger.Warnf("capture %d too large (%d bytes), skipping: %v", capture.ID, len(compressed), err)
-		return false
-	}
-
-	compressionRatio := (1 - float64(len(compressed))/float64(uncompressedBytes)) * 100
-	mp.logger.Debugf("Capture %d compressed and saved: %d bytes -> %d bytes (%.1f%% compression)", capture.ID, uncompressedBytes, len(compressed), compressionRatio)
-	return true
-}
-
-// getCompressedBytes returns the raw compressed bytes for a capture by ID.
-func (mp *metricsMonitor) getCompressedBytes(id int) ([]byte, bool) {
-	if mp.captureCache == nil {
-		return nil, false
-	}
-	data, err := mp.captureCache.Get(id)
-	if err != nil {
-		return nil, false
-	}
-	return data, true
-}
-
-// getCaptureByID decompresses and unmarshals a capture by ID.
-// Returns nil if the capture is not found or decompression fails.
-func (mp *metricsMonitor) getCaptureByID(id int) *ReqRespCapture {
-	if mp.captureCache == nil {
-		return nil
-	}
-	data, exists := mp.getCompressedBytes(id)
-	if !exists {
-		return nil
-	}
-
-	capture, err := decompressCapture(data)
-	if err != nil {
-		mp.logger.Warnf("failed to decompress capture %d: %v", id, err)
-		return nil
-	}
-
-	return capture
-}
-
-// getMetrics returns a copy of the current metrics with HasCapture resolved from cache.
-func (mp *metricsMonitor) getMetrics() []ActivityLogEntry {
-	mp.mu.RLock()
-	defer mp.mu.RUnlock()
-
-	result := mp.metrics.Slice()
-	if result == nil {
-		return []ActivityLogEntry{}
-	}
-	if mp.captureCache != nil {
-		for i := range result {
-			result[i].HasCapture = mp.captureCache.Has(result[i].ID)
-		}
-	}
-	return result
-}
-
-// getMetricsJSON returns metrics as JSON with HasCapture resolved from cache.
-func (mp *metricsMonitor) getMetricsJSON() ([]byte, error) {
-	mp.mu.RLock()
-	defer mp.mu.RUnlock()
-
-	result := mp.metrics.Slice()
-	if result == nil {
-		return json.Marshal([]ActivityLogEntry{})
-	}
-	if mp.captureCache != nil {
-		for i := range result {
-			result[i].HasCapture = mp.captureCache.Has(result[i].ID)
-		}
-	}
-	return json.Marshal(result)
-}
-
-// Capture field flags for controlling what is saved in ReqRespCapture.
-type captureFields uint
-
-const (
-	captureNone captureFields = 1 << iota
-	captureReqHeaders
-	captureReqBody
-	captureRespHeaders
-	captureRespBody
-)
-
-const (
-	captureReqAll  = captureReqHeaders | captureReqBody
-	captureRespAll = captureRespHeaders | captureRespBody
-	captureAll     = captureReqAll | captureRespAll
-)
-
-// wrapHandler wraps the proxy handler to extract token metrics.
-// captureFields controls what is saved in the ReqRespCapture using bitwise flags.
-// if wrapHandler returns an error it is safe to assume that no
-// data was sent to the client
-func (mp *metricsMonitor) wrapHandler(
-	modelID string,
-	writer gin.ResponseWriter,
-	request *http.Request,
-	captureFields captureFields,
-	next func(modelID string, w http.ResponseWriter, r *http.Request) error,
-) error {
-	// Capture request body and headers if captures enabled
-	var reqBody []byte
-	var reqHeaders map[string]string
-	if mp.enableCaptures && (captureFields&captureReqBody) != 0 {
-		if request.Body != nil {
-			var err error
-			reqBody, err = io.ReadAll(request.Body)
-			if err != nil {
-				return fmt.Errorf("failed to read request body for capture: %w", err)
-			}
-			request.Body.Close()
-			request.Body = io.NopCloser(bytes.NewBuffer(reqBody))
-		}
-	}
-	if mp.enableCaptures && (captureFields&captureReqHeaders) != 0 {
-		reqHeaders = make(map[string]string)
-		for key, values := range request.Header {
-			if len(values) > 0 {
-				reqHeaders[key] = values[0]
-			}
-		}
-		redactHeaders(reqHeaders)
-	}
-
-	recorder := newBodyCopier(writer)
-
-	// Filter Accept-Encoding to only include encodings we can decompress for metrics
-	if ae := request.Header.Get("Accept-Encoding"); ae != "" {
-		request.Header.Set("Accept-Encoding", filterAcceptEncoding(ae))
-	}
-
-	if err := next(modelID, recorder, request); err != nil {
-		return err
-	}
-
-	// after this point we have to assume that data was sent to the client
-	// and we can only log errors but not send them to clients
-
-	// Initialize default metrics - recorded for every request
-	tm := ActivityLogEntry{
-		Timestamp:       time.Now(),
-		Model:           modelID,
-		ReqPath:         request.URL.Path,
-		RespContentType: recorder.Header().Get("Content-Type"),
-		RespStatusCode:  recorder.Status(),
-		DurationMs:      int(time.Since(recorder.StartTime()).Milliseconds()),
-	}
-
-	if recorder.Status() != http.StatusOK {
-		mp.logger.Warnf("non-200 response, recording partial metrics: status=%d, path=%s", recorder.Status(), request.URL.Path)
-		tm.ID = mp.queueMetrics(tm)
-		mp.emitMetric(tm)
-		return nil
-	}
-
-	body := recorder.body.Bytes()
-	if len(body) == 0 {
-		mp.logger.Warn("metrics: empty body, recording minimal metrics")
-		tm.ID = mp.queueMetrics(tm)
-		mp.emitMetric(tm)
-		return nil
-	}
-
-	// Decompress if needed
-	if encoding := recorder.Header().Get("Content-Encoding"); encoding != "" {
-		var err error
-		body, err = decompressBody(body, encoding)
-		if err != nil {
-			mp.logger.Warnf("metrics: decompression failed: %v, path=%s, recording minimal metrics", err, request.URL.Path)
-			tm.ID = mp.queueMetrics(tm)
-			mp.emitMetric(tm)
-			return nil
-		}
-	}
-	if strings.Contains(recorder.Header().Get("Content-Type"), "text/event-stream") {
-		if parsed, err := processStreamingResponse(modelID, recorder.StartTime(), body); err != nil {
-			mp.logger.Warnf("error processing streaming response: %v, path=%s, recording minimal metrics", err, request.URL.Path)
-		} else {
-			tm.Tokens = parsed.Tokens
-			tm.DurationMs = parsed.DurationMs
-		}
-	} else {
-		if gjson.ValidBytes(body) {
-			parsed := gjson.ParseBytes(body)
-			usage := parsed.Get("usage")
-			timings := parsed.Get("timings")
-
-			// extract timings for infill - response is an array, timings are in the last element
-			// see #463
-			if strings.HasPrefix(request.URL.Path, "/infill") {
-				if arr := parsed.Array(); len(arr) > 0 {
-					timings = arr[len(arr)-1].Get("timings")
-				}
-			}
-
-			if usage.Exists() || timings.Exists() {
-				if parsedMetrics, err := parseMetrics(modelID, recorder.StartTime(), usage, timings); err != nil {
-					mp.logger.Warnf("error parsing metrics: %v, path=%s, recording minimal metrics", err, request.URL.Path)
-				} else {
-					tm.Tokens = parsedMetrics.Tokens
-					tm.DurationMs = parsedMetrics.DurationMs
-				}
-			}
-		} else {
-			mp.logger.Warnf("metrics: invalid JSON in response body path=%s, recording minimal metrics", request.URL.Path)
-		}
-	}
-
-	// Build capture if enabled and determine if it will be stored
-	var capture *ReqRespCapture
-	if mp.enableCaptures {
-		var respHeaders map[string]string
-		var respBody []byte
-		if (captureFields & captureRespHeaders) != 0 {
-			respHeaders = make(map[string]string)
-			for key, values := range recorder.Header() {
-				if len(values) > 0 {
-					respHeaders[key] = values[0]
-				}
-			}
-			redactHeaders(respHeaders)
-			delete(respHeaders, "Content-Encoding")
-		}
-		if (captureFields & captureRespBody) != 0 {
-			respBody = body
-		}
-		capture = &ReqRespCapture{
-			ReqPath:     request.URL.Path,
-			ReqHeaders:  reqHeaders,
-			ReqBody:     reqBody,
-			RespHeaders: respHeaders,
-			RespBody:    respBody,
-		}
-	}
-
-	metricID := mp.queueMetrics(tm)
-	tm.ID = metricID
-
-	// Store capture if enabled
-	if capture != nil {
-		capture.ID = metricID
-		if mp.addCapture(*capture) {
-			tm.HasCapture = true
-		}
-	}
-
-	mp.emitMetric(tm)
-
-	return nil
-}
-
-// usagePaths lists the JSON paths where a per-event usage object can live.
-// v1/chat/completions puts it at top-level "usage"; v1/responses nests under
-// "response.usage"; v1/messages emits it at "message.usage" on message_start
-// and at "usage" on message_delta.
-var usagePaths = []string{"usage", "response.usage", "message.usage"}
-
-// extractUsageTokens reads input/output/cached token counts from a usage
-// gjson.Result, handling the field-name differences across endpoints.
-// cached returns -1 when the field is absent. ok is true when at least one
-// field was present.
-func extractUsageTokens(usage gjson.Result) (input, output, cached int64, ok bool) {
-	cached = -1
-	if !usage.Exists() {
-		return
-	}
-
-	if v := usage.Get("prompt_tokens"); v.Exists() {
-		// v1/chat/completions
-		input = v.Int()
-		ok = true
-	} else if v := usage.Get("input_tokens"); v.Exists() {
-		// v1/messages, v1/responses
-		input = v.Int()
-		ok = true
-	}
-
-	if v := usage.Get("completion_tokens"); v.Exists() {
-		// v1/chat/completions
-		output = v.Int()
-		ok = true
-	} else if v := usage.Get("output_tokens"); v.Exists() {
-		// v1/messages, v1/responses
-		output = v.Int()
-		ok = true
-	}
-
-	if v := usage.Get("cache_read_input_tokens"); v.Exists() {
-		// v1/messages (Anthropic)
-		cached = v.Int()
-		ok = true
-	} else if v := usage.Get("input_tokens_details.cached_tokens"); v.Exists() {
-		// v1/responses (OpenAI Responses API)
-		cached = v.Int()
-		ok = true
-	} else if v := usage.Get("prompt_tokens_details.cached_tokens"); v.Exists() {
-		// v1/chat/completions (OpenAI cache hits)
-		cached = v.Int()
-		ok = true
-	}
-	return
-}
-
-func processStreamingResponse(modelID string, start time.Time, body []byte) (ActivityLogEntry, error) {
-	// Walk SSE "data:" lines forward, merging usage info from every event.
-	// Different endpoints split usage across events:
-	//   - v1/chat/completions: usage on the final chunk before [DONE]
-	//   - v1/responses:        usage on response.completed (response.usage)
-	//   - v1/messages:         input + cache on message_start (message.usage),
-	//                          output_tokens on message_delta (usage)
-	// We take the latest informative value per field so all three are covered.
-
-	var (
-		inputTokens, outputTokens int64
-		cachedTokens              int64 = -1
-		hasAny                    bool
-		timings                   gjson.Result
-	)
-
-	prefix := []byte("data:")
-	for offset := 0; offset < len(body); {
-		nl := bytes.IndexByte(body[offset:], '\n')
-		var line []byte
-		if nl == -1 {
-			line = body[offset:]
-			offset = len(body)
-		} else {
-			line = body[offset : offset+nl]
-			offset += nl + 1
-		}
-
-		line = bytes.TrimSpace(line)
-		if len(line) == 0 || !bytes.HasPrefix(line, prefix) {
-			continue
-		}
-		data := bytes.TrimSpace(line[len(prefix):])
-		if len(data) == 0 || bytes.Equal(data, []byte("[DONE]")) {
-			continue
-		}
-		if !gjson.ValidBytes(data) {
-			continue
-		}
-		parsed := gjson.ParseBytes(data)
-
-		for _, path := range usagePaths {
-			u := parsed.Get(path)
-			if !u.Exists() {
-				continue
-			}
-			i, o, c, ok := extractUsageTokens(u)
-			if !ok {
-				continue
-			}
-			hasAny = true
-			// Take the latest non-zero value so message_start's input_tokens
-			// is preserved when message_delta's usage omits it, and vice versa
-			// for output_tokens.
-			if i > 0 {
-				inputTokens = i
-			}
-			if o > 0 {
-				outputTokens = o
-			}
-			if c >= 0 {
-				cachedTokens = c
-			}
-		}
-		if t := parsed.Get("timings"); t.Exists() {
-			timings = t
-			hasAny = true
-		}
-	}
-
-	if !hasAny {
-		return ActivityLogEntry{}, fmt.Errorf("no valid JSON data found in stream")
-	}
-
-	return buildMetrics(modelID, start, inputTokens, outputTokens, cachedTokens, timings), nil
-}
-
-func parseMetrics(modelID string, start time.Time, usage, timings gjson.Result) (ActivityLogEntry, error) {
-	input, output, cached, _ := extractUsageTokens(usage)
-	return buildMetrics(modelID, start, input, output, cached, timings), nil
-}
-
-// buildMetrics composes an ActivityLogEntry from accumulated token counts and
-// optional llama-server timings (which override input/output and provide rates).
-func buildMetrics(modelID string, start time.Time, inputTokens, outputTokens, cachedTokens int64, timings gjson.Result) ActivityLogEntry {
-	wallDurationMs := int(time.Since(start).Milliseconds())
-	durationMs := wallDurationMs
-	tokensPerSecond := -1.0
-	promptPerSecond := -1.0
-
-	if timings.Exists() {
-		inputTokens = timings.Get("prompt_n").Int()
-		outputTokens = timings.Get("predicted_n").Int()
-		promptPerSecond = timings.Get("prompt_per_second").Float()
-		tokensPerSecond = timings.Get("predicted_per_second").Float()
-		timingsDurationMs := int(timings.Get("prompt_ms").Float() + timings.Get("predicted_ms").Float())
-		if timingsDurationMs > durationMs {
-			durationMs = timingsDurationMs
-		}
-		if cachedValue := timings.Get("cache_n"); cachedValue.Exists() {
-			cachedTokens = cachedValue.Int()
-		}
-	}
-
-	return ActivityLogEntry{
-		Timestamp: time.Now(),
-		Model:     modelID,
-		Tokens: TokenMetrics{
-			CachedTokens:    int(cachedTokens),
-			InputTokens:     int(inputTokens),
-			OutputTokens:    int(outputTokens),
-			PromptPerSecond: promptPerSecond,
-			TokensPerSecond: tokensPerSecond,
-		},
-		DurationMs: durationMs,
-	}
-}
-
-// decompressBody decompresses the body based on Content-Encoding header
-func decompressBody(body []byte, encoding string) ([]byte, error) {
-	switch strings.ToLower(strings.TrimSpace(encoding)) {
-	case "gzip":
-		reader, err := gzip.NewReader(bytes.NewReader(body))
-		if err != nil {
-			return nil, err
-		}
-		defer reader.Close()
-		return io.ReadAll(reader)
-	case "deflate":
-		reader := flate.NewReader(bytes.NewReader(body))
-		defer reader.Close()
-		return io.ReadAll(reader)
-	default:
-		return body, nil // Return as-is for unknown/no encoding
-	}
-}
-
-// responseBodyCopier records the response body and writes to the original response writer
-// while also capturing it in a buffer for later processing
-type responseBodyCopier struct {
-	gin.ResponseWriter
-	body  *bytes.Buffer
-	tee   io.Writer
-	start time.Time
-}
-
-func newBodyCopier(w gin.ResponseWriter) *responseBodyCopier {
-	bodyBuffer := &bytes.Buffer{}
-	return &responseBodyCopier{
-		ResponseWriter: w,
-		body:           bodyBuffer,
-		tee:            io.MultiWriter(w, bodyBuffer),
-		start:          time.Now(),
-	}
-}
-
-func (w *responseBodyCopier) Write(b []byte) (int, error) {
-	return w.tee.Write(b)
-}
-
-func (w *responseBodyCopier) WriteHeader(statusCode int) {
-	w.ResponseWriter.WriteHeader(statusCode)
-}
-
-func (w *responseBodyCopier) Header() http.Header {
-	return w.ResponseWriter.Header()
-}
-
-func (w *responseBodyCopier) StartTime() time.Time {
-	return w.start
-}
-
-// sensitiveHeaders lists headers that should be redacted in captures
-var sensitiveHeaders = map[string]bool{
-	"authorization":       true,
-	"proxy-authorization": true,
-	"cookie":              true,
-	"set-cookie":          true,
-	"x-api-key":           true,
-}
-
-// redactHeaders replaces sensitive header values in-place with "[REDACTED]"
-func redactHeaders(headers map[string]string) {
-	for key := range headers {
-		if sensitiveHeaders[strings.ToLower(key)] {
-			headers[key] = "[REDACTED]"
-		}
-	}
-}
-
-// filterAcceptEncoding filters the Accept-Encoding header to only include
-// encodings we can decompress (gzip, deflate). This respects the client's
-// preferences while ensuring we can parse response bodies for metrics.
-func filterAcceptEncoding(acceptEncoding string) string {
-	if acceptEncoding == "" {
-		return ""
-	}
-
-	supported := map[string]bool{"gzip": true, "deflate": true}
-	var filtered []string
-
-	for part := range strings.SplitSeq(acceptEncoding, ",") {
-		// Parse encoding and optional quality value (e.g., "gzip;q=1.0")
-		encoding, _, _ := strings.Cut(strings.TrimSpace(part), ";")
-		if supported[strings.ToLower(encoding)] {
-			filtered = append(filtered, strings.TrimSpace(part))
-		}
-	}
-
-	return strings.Join(filtered, ", ")
-}

proxy/peerproxy.go

@@ -1,144 +0,0 @@
-package proxy
-
-import (
-	"fmt"
-	"net"
-	"net/http"
-	"net/http/httputil"
-	"runtime"
-	"sort"
-	"strings"
-	"time"
-
-	"github.com/mostlygeek/llama-swap/internal/config"
-	"github.com/mostlygeek/llama-swap/internal/logmon"
-)
-
-type peerProxyMember struct {
-	peerID       string
-	reverseProxy *httputil.ReverseProxy
-	apiKey       string
-}
-
-type PeerProxy struct {
-	peers    config.PeerDictionaryConfig
-	proxyMap map[string]*peerProxyMember
-}
-
-func NewPeerProxy(peers config.PeerDictionaryConfig, proxyLogger *logmon.Monitor) (*PeerProxy, error) {
-	proxyMap := make(map[string]*peerProxyMember)
-
-	// Sort peer IDs for consistent iteration order
-	peerIDs := make([]string, 0, len(peers))
-	for peerID := range peers {
-		peerIDs = append(peerIDs, peerID)
-	}
-	sort.Strings(peerIDs)
-
-	for _, peerID := range peerIDs {
-		peer := peers[peerID]
-
-		// Create a transport with per-peer timeout configuration
-		peerTransport := &http.Transport{
-			Proxy: http.ProxyFromEnvironment,
-			DialContext: (&net.Dialer{
-				Timeout:   time.Duration(peer.Timeouts.Connect) * time.Second,
-				KeepAlive: time.Duration(peer.Timeouts.KeepAlive) * time.Second,
-			}).DialContext,
-			TLSHandshakeTimeout:   time.Duration(peer.Timeouts.TLSHandshake) * time.Second,
-			ResponseHeaderTimeout: time.Duration(peer.Timeouts.ResponseHeader) * time.Second,
-			ExpectContinueTimeout: time.Duration(peer.Timeouts.ExpectContinue) * time.Second,
-			ForceAttemptHTTP2:     true,
-			MaxIdleConns:          100,
-			MaxIdleConnsPerHost:   10,
-			IdleConnTimeout:       time.Duration(peer.Timeouts.IdleConn) * time.Second,
-		}
-
-		// Create reverse proxy for this peer
-		reverseProxy := httputil.NewSingleHostReverseProxy(peer.ProxyURL)
-		reverseProxy.Transport = peerTransport
-
-		// Wrap Director to set Host header for remote hosts (not localhost)
-		originalDirector := reverseProxy.Director
-		reverseProxy.Director = func(req *http.Request) {
-			originalDirector(req)
-			// Ensure Host header matches target URL for remote proxying
-			req.Host = req.URL.Host
-		}
-
-		reverseProxy.ModifyResponse = func(resp *http.Response) error {
-			if strings.Contains(strings.ToLower(resp.Header.Get("Content-Type")), "text/event-stream") {
-				resp.Header.Set("X-Accel-Buffering", "no")
-			}
-			return nil
-		}
-
-		reverseProxy.ErrorHandler = func(w http.ResponseWriter, r *http.Request, err error) {
-			proxyLogger.Warnf("peer %s: proxy error: %v", peerID, err)
-			errMsg := fmt.Sprintf("peer proxy error: %v", err)
-			if runtime.GOOS == "darwin" && strings.Contains(err.Error(), "connect: no route to host") {
-				errMsg += " (hint: on macOS, check System Settings > Privacy & Security > Local Network permissions)"
-			}
-			http.Error(w, errMsg, http.StatusBadGateway)
-		}
-
-		pp := &peerProxyMember{
-			peerID:       peerID,
-			reverseProxy: reverseProxy,
-			apiKey:       peer.ApiKey,
-		}
-
-		// Map each model to this peer's proxy
-		for _, modelID := range peer.Models {
-			if _, found := proxyMap[modelID]; found {
-				proxyLogger.Warnf("peer %s: model %s already mapped to another peer, skipping", peerID, modelID)
-				continue
-			}
-			proxyMap[modelID] = pp
-		}
-	}
-
-	return &PeerProxy{
-		peers:    peers,
-		proxyMap: proxyMap,
-	}, nil
-}
-
-func (p *PeerProxy) HasPeerModel(modelID string) bool {
-	_, found := p.proxyMap[modelID]
-	return found
-}
-
-// GetPeerFilters returns the filters for a peer model, or empty filters if not found
-func (p *PeerProxy) GetPeerFilters(modelID string) config.Filters {
-	pp, found := p.proxyMap[modelID]
-	if !found {
-		return config.Filters{}
-	}
-	// Get the peer config using the peerID
-	peer, found := p.peers[pp.peerID]
-	if !found {
-		return config.Filters{}
-	}
-	return peer.Filters
-}
-
-func (p *PeerProxy) ListPeers() config.PeerDictionaryConfig {
-	return p.peers
-}
-
-func (p *PeerProxy) ProxyRequest(model_id string, writer http.ResponseWriter, request *http.Request) error {
-	pp, found := p.proxyMap[model_id]
-	if !found {
-		return fmt.Errorf("no peer proxy found for model %s", model_id)
-	}
-
-	// Inject API key if configured for this peer
-	if pp.apiKey != "" {
-		request.Header.Set("Authorization", "Bearer "+pp.apiKey)
-		request.Header.Set("x-api-key", pp.apiKey)
-	}
-
-	pp.reverseProxy.ServeHTTP(writer, request)
-	return nil
-}