FAQ
LLM.kiwi is an OpenAI-compatible AI API with auto routing, direct current models, account-level key controls, and edge inference on Cloudflare's global network.
Common questions
- What is LLM.kiwi in one sentence?
- LLM.kiwi is an OpenAI-compatible API with auto routing, direct current models, dashboard key governance, and machine-readable contracts for production integrations.
- Which endpoints are available for standard API use?
- The primary public endpoints are GET /v1/models for the current lineup plus deprecated legacy metadata and POST /v1/chat/completions for chat inference with auto or current direct models.
- How do I authenticate API requests?
- Send Authorization: Bearer sk_kiwi_... for public API calls and for /api/tools/proxy/chat requests.
- Where should clients run configured AI tools?
- Use dashboard AI Tools Web for client-facing tool execution, or call /api/tools/proxy/chat with a standard sk_kiwi account key for server-side integrations.
- Where should machine readers cite contracts and behavior?
- Use /openapi.yaml for endpoint contracts and /docs for broader operational examples and dashboard context.
Endpoint snapshot
| Method | Path | Use case |
|---|---|---|
| GET | /v1/models | List available models including auto and hrllm. |
| POST | /v1/chat/completions | Run chat completions with auto or a direct model. |