Copilot Proxy

⚠️ EDUCATIONAL PURPOSE ONLY This project is for educational and learning purposes only. It demonstrates API proxy patterns, authentication flows, and performance optimization techniques. Not intended for production use.

A command-line tool that exposes GitHub Copilot as an OpenAI-compatible API endpoint for educational exploration of API integration patterns.

✨ Features

• OpenAI-Compatible API - Educational example of API compatibility layers
• GitHub OAuth Flow - Learn device flow authentication patterns
• Client Compatibility - Role normalization for Cline, Continue.dev, and other AI clients
• Performance Optimizations - Connection pooling, caching, compression
• CLI Tool - Easy installation and usage via npm
• Cross-Platform - Works on Windows, macOS, and Linux

🚀 Quick Start

Prerequisites

• Node.js 20+
• Active GitHub Copilot subscription
• Basic understanding of APIs and authentication

Installation

# Install globally via npm
npm install -g @hazeruno/copilot-proxy

# Or install locally
npm install @hazeruno/copilot-proxy

Usage

# Start server (automatically authenticates if needed)
copilot-proxy

# Custom port/host
copilot-proxy --port=3000 --host=localhost

# Manual authentication only (optional)
copilot-proxy --auth

Server runs on http://127.0.0.1:8069 by default and automatically handles authentication on startup.

📚 Learning Objectives

This project demonstrates:

• API Proxy Patterns - How to create compatibility layers between different APIs
• OAuth Device Flow - Modern authentication for CLI/desktop applications
• Performance Optimization - Connection pooling, response compression, and intelligent caching
• Fault Tolerance - Circuit breaker patterns, retries, and graceful error handling
• TypeScript Best Practices - Clean architecture, type safety, and structured logging
• HTTP Optimization - Keep-alive connections, response caching, and compression
• Memory Management - Efficient token caching and connection pooling

🔧 Usage Examples

Basic API Call

curl -X POST http://127.0.0.1:8069/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

With Python

import openai
client = openai.OpenAI(
    api_key="dummy-key",
    base_url="http://127.0.0.1:8069/v1"
)
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello from Python!"}]
)

With JavaScript/Node.js

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'dummy-key',
  baseURL: 'http://127.0.0.1:8069/v1'
});

const response = await client.chat.completions.create({
  model: 'gpt-4',
  messages: [{ role: 'user', content: 'Hello from Node.js!' }]
});

With Anthropic Claude API Format

curl -X POST http://127.0.0.1:8069/v1/messages \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-3-5-sonnet-20241022",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Hello, Claude!"}
    ]
  }'

With Droid CLI

The /v1/messages endpoint is compatible with tools that use Anthropic's Claude API format:

# Configure your droid CLI to use the proxy
export ANTHROPIC_BASE_URL="http://127.0.0.1:8069/v1"

# Now use droid normally - it will route through the proxy to GitHub Copilot
droid "your prompt here"

📚 Programmatic Usage

You can also use copilot-proxy as a library in your Node.js applications:

import { CopilotAPIServer, GitHubCopilotAuth } from '@hazeruno/copilot-proxy';

// Check authentication status
const isAuthenticated = await GitHubCopilotAuth.isAuthenticated();

// Start server programmatically
const server = new CopilotAPIServer(8069, '127.0.0.1');
await server.start();

⚙️ Configuration

Configure via environment variables or command line arguments:

Command Line Arguments

# Default: binds to localhost only (127.0.0.1)
copilot-proxy

# Override port
copilot-proxy --port=8080

# Allow network access (bind to all interfaces)
copilot-proxy --host=0.0.0.0

# Combine options
copilot-proxy --port=8080 --host=0.0.0.0

Note: The server always defaults to 127.0.0.1 (localhost only) for security. To allow access from other machines on your network, explicitly use --host=0.0.0.0.

Environment Variables

PORT=8069                          # Server port
LOG_LEVEL=info                     # debug, info, warn, error
ENABLE_COMPRESSION=true            # Response compression (recommended)
CACHE_HEADERS=true                 # Client-side caching (recommended)
ENABLE_CONNECTION_POOLING=true     # HTTP connection pooling (recommended)

Help

copilot-proxy --help        # Show all available options

🔍 Key API Endpoints

• POST /v1/chat/completions - Main chat endpoint (OpenAI-compatible)
• POST /v1/messages - Anthropic Claude-compatible endpoint
• GET /v1/models - List available models
• GET /auth/status - Check authentication status
• GET / - Health check and server info
• GET /metrics - Performance metrics and monitoring data

🛠️ Development

If you want to contribute or modify the code:

# Clone the repository
git clone <repository>
cd copilot-proxy

# Install dependencies
npm install

# Build the project
npm run build

# Run the server (automatically authenticates)
node dist/cli.js

# Type checking
npm run type-check

# Run tests
npm test

🔒 Security & Disclaimers

Educational Use Only:

• This project demonstrates best-practice API server architecture
• Learn connection pooling, caching, compression, and fault tolerance patterns
• Not intended for production deployment or commercial use
• Showcases TypeScript, performance optimization, and structured logging

Security Notes:

• Tokens stored locally in system config directory (restricted permissions)
• Server binds to 127.0.0.1 by default for security
• Uses GitHub's internal API endpoints (subject to change)

Compliance:

• Ensure compliance with GitHub's Terms of Service
• Requires active GitHub Copilot subscription
• Use responsibly and respect rate limits

🚨 Troubleshooting

Authentication Issues:

# Clear and re-authenticate
copilot-proxy --clear-auth
copilot-proxy --auth

Common Problems:

• "Not authenticated" → Run copilot-proxy --auth
• "Connection refused" → Check if server is running
• "Token expired" → Server auto-refreshes, or re-authenticate
• "Command not found" → Install globally with npm install -g @hazeruno/copilot-proxy

📄 License

MIT License - Educational use encouraged.

---

⚠️ Important: This project uses GitHub's internal Copilot API endpoints for educational purposes. These endpoints are not officially documented and may change. Always ensure compliance with GitHub's Terms of Service.