> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.jambonz.org/guides/features/bring-your-own-llm/moonshot-kimi/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.jambonz.org/_mcp/server. # Moonshot (Kimi) [Moonshot AI](https://www.moonshot.ai) serves its Kimi models (including the Kimi K2 mixture-of-experts family) behind an OpenAI-compatible API. The wire is OpenAI-compatible (chat completions, streaming, tools), so the integration is straightforward — jambonz talks to it via a base URL override. Two endpoints are available: * **Global** (default) — `https://api.moonshot.ai/v1` * **China** — `https://api.moonshot.cn/v1` ## Get credentials 1. Sign in at the global console ([platform.moonshot.ai](https://platform.moonshot.ai)) or the China console ([platform.moonshot.cn](https://platform.moonshot.cn)). 2. Open **API Keys** and create a key. 3. Copy the value — it is shown only once. ## Configure in jambonz In the portal: **Account → LLM Services → + Add LLM Service → Moonshot**. **`API Key`** `string` — required The key from Moonshot's console. --- **`Base URL`** `string` Defaults to `https://api.moonshot.ai/v1` (global). Set to `https://api.moonshot.cn/v1` for the China endpoint, or to your own proxy. --- Click **Test** to verify. ## Use in an agent verb ```js session.agent({ llm: { vendor: 'moonshot', model: 'kimi-k2-0711-preview', llmOptions: { systemPrompt: 'You are a helpful voice assistant.', }, }, stt: { vendor: 'deepgram', language: 'en-US' }, tts: { vendor: 'cartesia', voice: 'sonic-english' }, turnDetection: 'krisp', bargeIn: { enable: true }, actionHook: '/agent-complete', }).send(); ``` ## Available models | Model id | Notes | | ---------------------- | --------------------------------------------------------- | | `kimi-k2-0711-preview` | Kimi K2 — large MoE model, strong tool use | | `kimi-latest` | Alias tracking the latest Kimi chat model | | `moonshot-v1-8k` | Moonshot v1, 8k context — cheaper/faster for simple turns | | `moonshot-v1-32k` | Moonshot v1, 32k context | | `moonshot-v1-128k` | Moonshot v1, 128k context | Check `GET /v1/models` against your endpoint for the live list. ## Quirks & errors > **Note** > > **Global vs China**: keep `vendor: 'moonshot'` constant — only the `Base URL` (in the portal credential) differs between the two regions. Keys are not interchangeable across endpoints. > **Warning** > > **`401`** — key revoked, copied with whitespace, or used against the wrong region's endpoint. Regenerate in the Moonshot console and confirm the `Base URL` matches the console you created the key in. > Configure jambonz to use Moonshot AI's Kimi models over its OpenAI-compatible API.