OpenClaw Setup

OpenClaw is an open-source AI agent platform that can use Clauddy as its model provider.

What is OpenClaw?

A local AI assistant framework supporting:

• Multi-channel messaging (Telegram, Discord, WhatsApp, etc.)
• Tool use (browser, filesystem, shell, etc.)
• Memory systems and extensible skills

Install

npm install -g openclaw

Configure Clauddy as Provider

Edit ~/.openclaw/openclaw.json and add Clauddy under models.providers:

Anthropic Models (Claude)

{
  "models": {
    "providers": {
      "clauddy": {
        "baseUrl": "https://clauddy.com",
        "apiKey": "sk-your-clauddy-token",
        "auth": "api-key",
        "api": "anthropic-messages",
        "authHeader": true,
        "models": [
          {
            "id": "claude-sonnet-4-6",
            "name": "claude-sonnet-4-6",
            "api": "anthropic-messages",
            "reasoning": true,
            "input": ["text", "image"],
            "contextWindow": 200000,
            "maxTokens": 8192,
            "compat": {
              "supportsReasoningEffort": true,
              "maxTokensField": "max_tokens"
            }
          },
          {
            "id": "claude-opus-4-6",
            "name": "claude-opus-4-6",
            "api": "anthropic-messages",
            "reasoning": true,
            "input": ["text", "image"],
            "contextWindow": 200000,
            "maxTokens": 8192,
            "compat": {
              "supportsReasoningEffort": true,
              "maxTokensField": "max_tokens"
            }
          }
        ]
      }
    }
  }
}

OpenAI Models (GPT / Codex)

{
  "models": {
    "providers": {
      "clauddy-openai": {
        "baseUrl": "https://clauddy.com/v1",
        "apiKey": "sk-your-clauddy-token",
        "api": "openai-responses",
        "models": [
          {
            "id": "gpt-5.3-codex",
            "name": "gpt-5.3-codex",
            "api": "openai-responses",
            "reasoning": true,
            "input": ["text"],
            "contextWindow": 200000,
            "maxTokens": 16384
          }
        ]
      }
    }
  }
}

Google Models (Gemini)

{
  "models": {
    "providers": {
      "clauddy-gemini": {
        "baseUrl": "https://clauddy.com/v1beta",
        "apiKey": "sk-your-clauddy-token",
        "auth": "api-key",
        "api": "google-generative-ai",
        "models": [
          {
            "id": "gemini-3.1-pro-preview",
            "name": "gemini-3.1-pro-preview",
            "api": "google-generative-ai",
            "reasoning": true,
            "input": ["text", "image"],
            "contextWindow": 1000000,
            "maxTokens": 65536
          }
        ]
      }
    }
  }
}

Field Reference

Field	Description
baseUrl	https://clauddy.com for Anthropic; https://clauddy.com/v1 for OpenAI; https://clauddy.com/v1beta for Gemini
apiKey	Your Clauddy API token
api	Protocol: anthropic-messages (Claude), openai-responses (GPT), google-generative-ai (Gemini)
auth	Set to api-key
authHeader	Set true for Anthropic models
models	Model list with id, name, capabilities

Using Models

Reference models as provider/model-id:

clauddy/claude-sonnet-4-6
clauddy/claude-opus-4-6
clauddy-openai/gpt-5.3-codex
clauddy-gemini/gemini-3.1-pro-preview

Set a default model:

{
  "models": {
    "default": "clauddy/claude-sonnet-4-6"
  }
}

Apply Changes

After editing openclaw.json, restart the gateway to apply:

openclaw gateway restart

TIP

OpenClaw supports hot-switching models. Use /model in chat to switch between configured models anytime.

Troubleshooting

Model switching not working

If you configured new models but OpenClaw can't recognize or switch to them, the model cache may be stale.

Fix:

1. Delete the cached models file:

rm ~/.openclaw/agents/main/agent/models.json

2. Restart the gateway:

openclaw gateway restart

OpenClaw will reload model configs from openclaw.json on restart.

Connection failed

• Verify baseUrl is correct (Anthropic: https://clauddy.com, OpenAI: https://clauddy.com/v1, Gemini: https://clauddy.com/v1beta)
• Verify apiKey is valid and the token group includes the target models
• Run openclaw gateway restart to reload config

More at OpenClaw docs.