Kimi Integration

**Quick judgment** High‑quality, focused integration guide for connecting Moonshot AI (Kimi K2) and Kimi Code models to Clawdbot. It should work well for most users who follow the steps exactly, and it supports both single‑provider and dual‑provider setups with aliases. **What it does well** - Clearly distinguishes Moonshot (general Kimi K2.x) vs Kimi Code (coding model) with separate keys and endpoints. - Provides ready‑to‑use `clawdbot.json` provider/model configs and environment variable patterns. - Shows how to set defaults and aliases so users can easily switch between `moonshot/kimi-k2.5` and `kimicode/kimi-for-coding`. - Adds practical troubleshooting commands and a direct curl example to test connectivity and authentication. **Key risks / limitations** - **API drift risk**: If Moonshot or Kimi Code change their endpoints, authentication scheme, or model IDs, some examples may become outdated and require adjustment. - **Key handling**: While it suggests env vars, users still must ensure keys are not committed to version control or exposed in logs. - **Provider‑specific scope**: The instructions are tailored to Clawdbot and OpenAI‑compatible APIs; they don’t cover broader DevOps concerns (rate limits, cost monitoring, or regional network issues). **Recommended scenarios** - You already run Clawdbot and want to add Kimi K2.5 as your main LLM. - You want a specialized coding model (Kimi for Coding) alongside a general model and need simple model switching via `/model` aliases. - You’re debugging missing Kimi/Moonshot models in Clawdbot or authentication/connection errors to those providers.