Reference

API reference

Endpoint docs with examples — chat, embeddings, tools, and local runners.

11 docs

Send multi-turn messages; get assistant replies. Use **server-side** only.

Turn text into vectors for semantic search and RAG.

Production apps need retries, backoff, and clear user errors.

Send conversations to Claude models. Keys stay **server-side**.

Let the model return **structured tool calls**; your server executes functions and sends results back.

Let Claude request **tools** with structured inputs — your code executes and returns results.

Run open-weight models on your machine; hit **localhost** from your app.

Standard **embed → search → pack → generate** loop for question answering over your docs.

Multimodal Google AI endpoint — text in, text out (images optional in newer models).

Return **JSON matching a schema** — reliable parsing for agents and APIs.

Same **chat completions** endpoint with image parts in messages.