steveandClaude Opus 5
02cd561eaf
Adds the `qwen` built-in provider and the `qwen://` DSN scheme, keyed by QWEN_API_KEY and defaulting to Model Studio's international host. Like kimi (ADR-0026) it is `provider/openai` pointed elsewhere — no new client. Model Studio serves the same models over two protocols, so the real decision was which wire format to speak. ADR-0027 records why it is the OpenAI one: down the anthropic client `ReasoningEffort` is ignored by design, structured output rides the first-party `output_config.format` mechanism the shim does not implement, and cached-token accounting reads Anthropic-only usage fields. Each of those fails silently rather than loudly, which is what makes the choice worth writing down. The shim stays reachable ad hoc via an `anthropic://` DSN. The kimi and qwen DSN factories were byte-identical, so they now share one `openaiCompatScheme` helper: the "credential comes from the DSN token, and the missing-key hint names LLM_<NAME>" rules hold by construction instead of by copy. Tests are hermetic and break-checked (all six fail on a deliberate mutation), including the reverse credential leak — a visible QWEN_API_KEY must not authenticate the openai built-in — and reasoning_effort asserted on the wire body, which is the ADR's load-bearing claim. Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
3.1 KiB
3.1 KiB
Architecture Decision Records
One decision per file, append-only; supersede rather than rewrite.
| ADR | Title | Status |
|---|---|---|
| 0001 | Package layout — canonical types in leaf llm, root re-exports |
Accepted |
| 0002 | Canonical message/content model | Accepted |
| 0003 | Parse grammar — verbatim ids, inline alias expansion, chains | Accepted |
| 0004 | LLM_* env-DSN provider definitions (go-llm parity + eager load) | Accepted |
| 0005 | Provider interface and capabilities model | Accepted |
| 0006 | Model health tracking and backoff | Accepted |
| 0007 | Dependency policy — stdlib-first, hand-rolled REST clients | Accepted |
| 0008 | Failover-chain execution semantics | Accepted |
| 0009 | Multimodal strategy — normalize per target, enforce at provider | Accepted |
| 0010 | Tools and structured output — canonical shape, native mappings | Accepted |
| 0011 | Google provider on the official Gen AI SDK | Accepted |
| 0012 | Agent run loop | Accepted |
| 0013 | Skill model — additive instruction+tool bundles | Accepted |
| 0014 | Conversion-driven extensions (resolvers, typed tools, hooks, ops controls) | Accepted |
| 0015 | llama-swap provider — reuse openai for chat, tailored management + image | Accepted |
| 0016 | imagegen — a canonical text-to-image interface | Accepted |
| 0017 | audio — canonical speech synthesis + transcription interfaces | Accepted |
| 0018 | imagegen.Editor — image-to-image as a separate optional interface | Accepted |
| 0019 | videogen — canonical video-generation surface | Accepted |
| 0020 | Upstream-passthrough media surfaces (mask, upscale, background removal, interpolation, diarization, meshgen) | Accepted |
| 0021 | musicgen — blocking Generate over an async job queue | Accepted |
| 0022 | embeddings + rerank interface | Accepted |
| 0023 | Wave-3 image + document surfaces (segmentation, colorize, face restore, OCR) | Accepted |
| 0024 | Wave-3 audio surfaces (stems, SFX, speech enhance, voice clone, translate) | Accepted |
| 0025 | Wave-3 video surfaces (lipsync, video matte, video upscale, chain jobs) | Accepted |
| 0026 | Kimi (Moonshot AI) built-in provider — reuse openai client, KIMI_API_KEY | Accepted |
| 0027 | Qwen (Alibaba) built-in provider — OpenAI-compat, not Model Studio's Anthropic-compat endpoint | Accepted |