Claude Code Router

Claude Code Router is a powerful routing tool that allows you to build coding infrastructure based on Claude Code, while letting you decide how to interact with models and enjoy updates from Anthropic.

GitHub: musistudio/claude-code-router
Stars: 10.3k+
License: MIT

🚀 Core Features

Multi-Model Intelligent Routing

• Smart Routing: Automatically selects appropriate models based on task type
• Cost Optimization: Uses smaller local models for background tasks
• Long Context Handling: Dedicated models for long context tasks (>60K tokens)
• Reasoning Enhancement: Specialized models for reasoning-intensive tasks

Multi-Provider Support

• OpenAI: Complete OpenAI API support
• Anthropic: Native Claude API support
• DeepSeek: Adapted for DeepSeek API
• Gemini: Google Gemini API support
• OpenRouter: Multi-provider routing support
• Groq: High-performance inference support

Transformer System

• maxtoken: Sets specific max_tokens values
• tooluse: Optimizes tool usage via tool_choice
• reasoning: Processes reasoning_content field
• sampling: Handles temperature, top_p, top_k, and other sampling parameters
• enhancetool: Adds error tolerance layer to tool call parameters

📦 Installation and Configuration

Quick Start

# Install
bunx @musistudio/claude-code-router@latest

# Create config directory
mkdir -p $HOME/.claude-code-router

# Copy example config
cp config.example.json $HOME/.claude-code-router/config.json

Basic Configuration

{
  "log": true,
  "OPENAI_API_KEY": "your-api-key",
  "OPENAI_BASE_URL": "https://api.deepseek.com",
  "OPENAI_MODEL": "deepseek-chat",
  "router": {
    "default": "openrouter,anthropic/claude-3.5-sonnet",
    "background": "openrouter,anthropic/claude-3-haiku",
    "think": "openrouter,anthropic/claude-3.5-sonnet",
    "longContext": "openrouter,anthropic/claude-3.5-sonnet",
    "longContextThreshold": 60000
  }
}

🔧 Routing Configuration

Default Routing Scenarios

{
  "router": {
    "default": "openrouter,anthropic/claude-3.5-sonnet",
    "background": "openrouter,anthropic/claude-3-haiku",
    "think": "openrouter,anthropic/claude-3.5-sonnet",
    "longContext": "openrouter,anthropic/claude-3.5-sonnet",
    "webSearch": "openrouter,anthropic/claude-3.5-sonnet:online"
  }
}

Dynamic Model Switching

Use the /model command in Claude Code to dynamically switch models:

/model openrouter,anthropic/claude-3.5-sonnet

⚡ Advanced Features

Custom Router

Create custom routing logic:

// $HOME/.claude-code-router/custom-router.js
module.exports = async function router(req, config) {
  const userMessage = req.body.messages.find((m) => m.role === "user")?.content;

  if (userMessage && userMessage.includes("explain this code")) {
    return "openrouter,anthropic/claude-3.5-sonnet";
  }

  return null; // Fallback to default router
};

Subagent Routing

Specify a particular model at the beginning of subagent prompts:

<CCR-SUBAGENT-MODEL>openrouter,anthropic/claude-3.5-sonnet</CCR-SUBAGENT-MODEL>
Please help me analyze this code snippet for potential optimizations...

GitHub Actions Integration

Integrate Claude Code Router into CI/CD pipelines:

name: Claude Code

on:
  issue_comment:
    types: [created]

jobs:
  claude:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout repository
        uses: actions/checkout@v4

      - name: Prepare Environment
        run: |
          curl -fsSL https://bun.sh/install | bash
          mkdir -p $HOME/.claude-code-router
          cat << 'EOF' > $HOME/.claude-code-router/config.json
          {
            "log": true,
            "OPENAI_API_KEY": "${{ secrets.OPENAI_API_KEY }}",
            "OPENAI_BASE_URL": "https://api.deepseek.com",
            "OPENAI_MODEL": "deepseek-chat"
          }
          EOF

      - name: Start Claude Code Router
        run: |
          nohup ~/.bun/bin/bunx @musistudio/claude-code-router@1.0.8 start &

      - name: Run Claude Code
        uses: anthropics/claude-code-action@beta
        env:
          ANTHROPIC_BASE_URL: http://localhost:3456
        with:
          anthropic_api_key: "any-string-is-ok"

🎯 Use Cases

1. Cost Optimization

• Use smaller models for simple tasks
• Run tasks during off-peak hours
• Reduce API call costs through intelligent routing

2. Performance Optimization

• Use dedicated models for long context tasks
• Use high-performance models for reasoning tasks
• Use lightweight models for background tasks

3. Multi-Provider Support

• Avoid single provider limitations
• Leverage advantages of different providers
• Improve service availability

💡 Best Practices

Configuration Recommendations

1. Set appropriate thresholds: Adjust longContextThreshold based on project requirements
2. Monitor usage: Enable logging to monitor model usage
3. Regular updates: Keep tools and configurations up to date
4. Test configurations: Thoroughly test configurations before production

Security Considerations

1. API key management: Store sensitive information using environment variables
2. Access control: Limit router access permissions
3. Log security: Avoid logging sensitive information

🤝 Community Support

Sponsorship Support

The project supports multiple sponsorship methods:

• Ko-fi
• PayPal
• Alipay
• WeChat Pay

Contribution Guidelines

Welcome contributions through:

1. Submit Issues to report problems
2. Create Pull Requests to contribute code
3. Share usage experiences and best practices
4. Sponsor project development

📝 Summary

Claude Code Router provides powerful routing capabilities for Claude Code, enabling you to:

• Flexibly select models: Choose appropriate AI models based on task requirements
• Optimize costs: Reduce API call costs through intelligent routing
• Improve performance: Select optimal models for different task types
• Enhance availability: Support multiple AI providers to avoid single points of failure

This tool is particularly suitable for teams and individual developers who need to handle large amounts of AI interactions, significantly improving development efficiency and cost-effectiveness.

Claude Code Router is an indispensable tool in the Claude Code ecosystem, providing developers with flexible, efficient, and economical AI model routing solutions.