How to Integrate Moonshot AI Models in Openclaw

Table of contents: [Show]

Want to use Kimi K2.5 in Openclaw? Moonshot AI (Kimi Open Platform) offers powerful models with OpenAI-compatible endpoints. This guide shows you how to integrate Moonshot AI models into your Openclaw setup in under 5 minutes.

Quick Start

Moonshot AI provides both global and China endpoints. Choose the one that fits your location. You'll need a MOONSHOT_API_KEY from the Moonshot platform.

Interactive Setup

Run the onboarding wizard to configure Moonshot AI:

For global endpoint:

openclaw onboard --auth-choice moonshot-api-key

For China endpoint:

openclaw onboard --auth-choice moonshot-api-key-cn

Manual Configuration

If you prefer manual setup, add this to your ~/.openclaw/openclaw.json:

{
  "env": { "MOONSHOT_API_KEY": "sk-..." },
  "agents": {
    "defaults": {
      "model": { "primary": "moonshot/kimi-k2.5" }
    }
  },
  "models": {
    "mode": "merge",
    "providers": {
      "moonshot": {
        "baseUrl": "https://api.moonshot.ai/v1",
        "apiKey": "${MOONSHOT_API_KEY}",
        "api": "openai-completions",
        "models": [
          {
            "id": "kimi-k2.5",
            "name": "Kimi K2.5",
            "reasoning": false,
            "contextWindow": 256000,
            "maxTokens": 8192
          }
        ]
      }
    }
  }
}

For more deployment options, check out our guide on Deploy Flyio Openclaw.

Available Models

All Moonshot models must be prefixed with moonshot/:

moonshot/kimi-k2.5
moonshot/kimi-k2-0905-preview
moonshot/kimi-k2-turbo-preview
moonshot/kimi-k2-thinking
moonshot/kimi-k2-thinking-turbo

Native Thinking Mode

Moonshot Kimi supports binary native thinking. Openclaw maps runtime /think levels to toggle this feature. To permanently disable thinking for a model:

"params": {
  "thinking": { "type": "disabled" }
}

Note: When thinking is enabled, tool choice must be set to auto or none. Openclaw automatically normalizes incompatible tool choices.

If you're new to Openclaw, you might also want to read about Elevated Mode Openclaw for advanced permissions.

Example Usage

Once configured, use Moonshot models in your prompts:

@moonshot/kimi-k2.5 Explain quantum computing in simple terms

Troubleshooting

Invalid API key: Verify your MOONSHOT_API_KEY is correctly set in environment or config
Endpoint errors: Use moonshot-api-key-cn for China region
Tool choice errors: When using thinking mode, ensure tool choice is auto
Model not found: Always prefix model IDs with moonshot/

Best Practices

Use kimi-k2.5 for general tasks with 256K context window
Enable thinking mode for complex reasoning tasks
Store API keys in environment variables, not hardcoded in config
Test with openclaw status to verify model connectivity