A FastAPI proxy server that seamlessly turns GitHub Copilot's chat completion/embeddings capabilities into OpenAI compatible API service.

🚀 Advanced Integration

• Seamless GitHub Copilot chat completion API proxy
• Real-time streaming response support
• High-performance request handling

🔐 Security & Reliability

• Secure authentication middleware
• Automatic token management and refresh
• Built-in CORS support for web applications

💻 Universal Compatibility

• Cross-platform support (Windows and Unix-based systems)
• Docker containerization ready
• Flexible deployment options

• Python 3.10+
• pip (Python package manager)
• GitHub Copilot subscription
• GitHub authentication token

1. Clone the repository:

git clone https://github.com/yuchanns/copilot-openai-api.git cd copilot_provider

2. Install dependencies:

# Install PDM first if you haven't pip install -U pdm # Install project dependencies using PDM pdm install --prod

1. Set up environment variables:

# Windows set COPILOT_TOKEN=your_access_token_here set COPILOT_SERVER_PORT=9191 # Optional: Server port (default: 9191) set COPILOT_SERVER_WORKERS=4 # Optional: Number of workers (default: min(CPU_COUNT, 4)) # Unix/Linux/macOS export COPILOT_TOKEN=your_access_token_here export COPILOT_SERVER_PORT=9191 # Optional: Server port (default: 9191) export COPILOT_SERVER_WORKERS=4 # Optional: Number of workers (default: min(CPU_COUNT, 4))

📝 Note:

• COPILOT_TOKEN: Required for authentication. If not set, a random token will be generated.
• COPILOT_SERVER_PORT: Optional. Controls which port the server listens on.
• COPILOT_SERVER_WORKERS: Optional. Controls the number of worker processes.

2. Configure GitHub Copilot:
   • Windows users: Check %LOCALAPPDATA%/github-copilot/
   • Unix/Linux/macOS users: Check ~/.config/github-copilot/

Required configuration files:

• apps.json or hosts.json (containing GitHub OAuth token)
• token.json (will be created automatically)

💡 How to get a valid Github Copilot configuration?

Choose any of these official GitHub Copilot plugins:

• GitHub Copilot for VS Code
• GitHub Copilot for Visual Studio
• GitHub Copilot for Vim
• GitHub Copilot for JetBrains IDEs

After installing and signing in, configuration files will be automatically created in your system's config directory.

Choose between local or Docker deployment:

Start the server with:

pdm dev

Launch the containerized version:

# Unix/Linux/macOS docker run --rm -p 9191:9191 \ -v ~/.config/github-copilot:/home/appuser/.config/github-copilot \ ghcr.io/yuchanns/copilot-openai-api # Windows docker run --rm -p 9191:9191 ^ -v %LOCALAPPDATA%/github-copilot:/home/appuser/.config/github-copilot ^ ghcr.io/yuchanns/copilot-openai-api

The Docker setup:

• Maps port 9191 to your host
• Mounts your Copilot configuration
• Provides identical functionality to local deployment

Access the chat completion endpoint:

curl -X POST http://localhost:9191/v1/chat/completions \ -H "Authorization: Bearer your_access_token_here" \ -H "Content-Type: application/json" \ -d '{  "messages": [{"role": "user", "content": "Hello, Copilot!"}]  }'

Access the embeddings endpoint:

curl -X POST http://localhost:9191/v1/embeddings \ -H "Authorization: Bearer your_access_token_here" \ -H "Content-Type: application/json" \ -d '{  "model": "copilot-text-embedding-ada-002",  "input": ["The quick brown fox", "Jumped over the lazy dog"]  }'

Proxies requests to GitHub Copilot's Models API to list available models.

Required Headers:

• Authorization: Bearer <your_access_token>

Response:

• JSON response containing a list of available models from GitHub Copilot
• Compatible with OpenAI API format

Example:

curl http://localhost:9191/v1/models \ -H "Authorization: Bearer your_access_token_here"

Get details for a specific model.

Required Headers:

• Authorization: Bearer <your_access_token>

Path Parameters:

• model: Model ID (e.g., gpt-5.1, claude-sonnet-4)

Response:

• JSON response with model details in OpenAI format

Example:

curl http://localhost:9191/v1/models/gpt-5.1 \ -H "Authorization: Bearer your_access_token_here"

Proxies requests to GitHub Copilot's Completions API.

Required Headers:

• Authorization: Bearer <your_access_token>
• Content-Type: application/json

Request Body:

• Follow GitHub Copilot chat completion API format

Response:

• Streams responses directly from GitHub Copilot's API
• Supports both streaming and non-streaming modes

Proxies requests to Github Copilot's Embeddings API.

Required Headers:

• Authorization: Bearer <your_access_token>
• Content-Type: application/json

Request Body:

• Follow GitHub Copilot embeddings API format

Response:

• JSON response from GitHub Copilot's embeddings API

Secure your endpoints:

1. Set COPILOT_TOKEN in your environment
2. Include in request headers:

   Authorization: Bearer your_access_token_here 

The server provides clear error responses:

• 401: Missing/invalid authorization header
• 403: Invalid access token
• Other errors are propagated from GitHub Copilot API

• Configure CORS for your specific domains (default: *)
• Safeguard your COPILOT_TOKEN and GitHub OAuth token
• Built-in token management with concurrent access protection

Licensed under the Apache License 2.0 - see the LICENSE file for details.