claudelisted

# Claude on RunAPI Use Anthropic Messages at `https://runapi.ai` with `POST /v1/messages` as the primary protocol. ## Primary protocol recipe ### Authenticate Set `ANTHROPIC_API_KEY` to a RunAPI API key and `ANTHROPIC_BASE_URL` to `https://runapi.ai`. The SDK sends `x-api-key`; Bearer authentication is also accepted. ### Send request ```python import anthropic client = anthropic.Anthropic(api_key="YOUR_RUNAPI_TOKEN", base_url="https://runapi.ai") message = client.messages.create( model="claude-sonnet-5", max_tokens=1024, messages=[{"role": "user", "content": "Explain this decision."}], ) print(message.content[0].text) print(message.usage) ``` `max_tokens` is required. For long output, use `client.messages.stream(...)` and consume `stream.text_stream` through the terminal event. Add image blocks, function tools, reasoning controls, or token counting only when the current RunAPI contract verifies them for the selected exact model. ### Verify result Require final text, `stop_reason`, and authoritative `usage`. A stream is complete only after terminal Usage and `message_stop`. Token counting estimates input only; actual response Usage remains authoritative. ### Stop boundaries Correct a rejected shape once using the structured error. Retry transport once only before any response or Usage and when replay is safe. Record a terminal error and stop without changing model or protocol. ## Compatibility protocols Load [compatibility protocols](references/compati