deepseek-harness/docs/cookbook/adding-an-llm-adapter.md

Cookbook: adding an LLM adapter

How to connect a new model provider. Reference implementations: packages/llm-deepseek (hand-rolled HTTP/SSE) and packages/llm-pi-ai (wrapping an LLM library). Read the StreamChunk doc in packages/llm/src/types.ts first — it records the protocol conventions both adapters were verified against.

The shape

class MyAdapter extends LlmAdapter {
  async * stream(options: GenerateOptions): AsyncIterable<StreamChunk> { … }
}

export const name = 'llm-myprovider'
export const inject = ['llm']
export const Config: z<Config> = z.object({ apiKey: z.string(), … })

export function apply(ctx: Context, config: Config) {
  ctx.llm.registerAdapter(['model-a', 'model-b'], new MyAdapter(…))
}

Registration is effect-based (HMR-safe); one adapter per model name — duplicates throw. Secrets are cordis-native: schemastery Config with env fallbacks, fed from cordis.yml via !!js process.env.MY_KEY. Never read ad-hoc key files in code.

Protocol obligations (the contract two implementations verified)

• Emit usage BEFORE finish; emit NOTHING after finish. The robust way: buffer finish/usage until the provider's end-of-stream marker, then flush (handles providers that send trailing usage-only chunks).
• Tool-call arguments are RAW JSON strings end-to-end; stream fragments as argumentsDelta. If your provider hands back parsed objects, re-stringify at block-end.
• Allocate block indexes in first-seen stream order; reuse the index for every delta of the same block.
• Errors have exactly two sanctioned paths: THROW from stream() (transport and protocol failures — use LlmError with a stable code), or end the stream with finish {kind: 'error' | 'aborted'} (provider in-band failures). Consumers handle both; pick per failure class and document it.
• Honor options.signal (pass it to fetch / your SDK).
• prefill and other unsupported GenerateOptions fields: throw LlmError(..., 'UNSUPPORTED') rather than silently dropping.

Provider-specific request knobs (thinking modes, effort levels) belong in the ADAPTER's Config, not in GenerateOptions — the core vocabulary stays provider-neutral.

Structure that worked

Split the adapter into testable stages (llm-deepseek's layout): wire types (types.ts, coverage-exempt) → request serializer → SSE/transport parser → chunk-translation state machine → a thin adapter class wiring them. Each stage gets its own unit suite.

Testing

• Unit: mock the provider, not the harness. A scripted node:http server speaking the provider's wire format covers happy paths, every error status, malformed payloads, premature closes, and aborts — no network, and it drives the 100% per-file coverage gate. Works for SDK-backed adapters too (point the SDK's baseURL at the mock).
• Hostile framing tests. Split stream payloads at arbitrary byte positions (including mid-UTF-8) — real networks do.
• E2E: tests/*.e2e.ts under pnpm run test:e2e, gated with describe.skipIf(!process.env.MY_KEY) so CI (no secrets) stays green. Cover each model × each provider mode you map (thinking on/off, effort levels), a tool-call round trip INCLUDING the follow-up turn with results in history, and loose assertions only (substring/structure, bounded maxTokens — real models are nondeterministic).
• Register the e2e file pattern in knip.json (per-workspace entry override) or knip flags it unused.