Tiat/context7-server

Context7 OpenAI-Compatible API Server

This is an OpenAI-compatible API server that integrates with Context7 MCP (Model Context Protocol) to provide up-to-date documentation for various libraries and frameworks.

Features

• —OpenAI-Compatible API: Drop-in replacement for OpenAI's chat completions endpoint
• —Context7 Integration: Automatically fetches relevant documentation using Context7 MCP
• —Real-time Documentation: Get the latest documentation for popular libraries
• —Easy Integration: Use with any OpenAI-compatible client

API Endpoints

Chat Completions (OpenAI Compatible)

POST /v1/chat/completions

Send OpenAI-compatible chat completion requests:

{
  "model": "context7-mcp",
  "messages": [
    {
      "role": "user", 
      "content": "How do I create a FastAPI app?"
    }
  ]
}

MCP Server Endpoint (for ChatGPT Desktop)

POST /sse

Server-Sent Events endpoint for Model Context Protocol integration with ChatGPT Desktop.

Health Check

GET /health

Returns the server health status.

Usage

You can use this server as a drop-in replacement for OpenAI's API in your applications. Just point your OpenAI client to this server's URL.

Example with Python OpenAI client:

import openai

client = openai.OpenAI(
    api_key="dummy-key",  # Not required for this server
    base_url="https://your-space-name.hf.space/v1"
)

response = client.chat.completions.create(
    model="context7-mcp",
    messages=[
        {"role": "user", "content": "How do I use React hooks?"}
    ]
)

print(response.choices[0].message.content)

Supported Libraries

The server automatically detects mentions of popular libraries and frameworks in your queries and fetches relevant documentation from Context7, including:

• —React, Next.js, Vue, Angular
• —FastAPI, Django, Express
• —MongoDB, PostgreSQL
• —And many more...

How It Works

1. 1.Query Processing: Extracts library names from user queries
2. 2.Context7 Integration: Uses Context7 MCP to resolve library IDs and fetch documentation
3. 3.Response Generation: Combines the documentation with the user's query to provide helpful responses
4. 4.OpenAI Compatibility: Returns responses in OpenAI's standard format

Development

To run locally:

# Build the Docker image
docker build -t context7-server .

# Run the container
docker run -p 7860:7860 context7-server

Technologies Used

• —FastAPI: Web framework for the API server
• —Context7 MCP: For fetching up-to-date documentation
• —Docker: For containerization and deployment
• —Hugging Face Spaces: For hosting and deployment

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

ChatGPT Desktop MCP Configuration

To use this server with ChatGPT Desktop's Model Context Protocol (MCP) feature:

1. Find Your Space URL

Once your Hugging Face Space is deployed, you'll have a URL like:

https://your-username-context7-server.hf.space

2. Configure ChatGPT Desktop

In ChatGPT Desktop, when prompted for the MCP server URL, enter:

https://your-username-context7-server.hf.space/sse

Important:

• —Make sure to use the /sse endpoint for ChatGPT Desktop MCP integration
• —The URL format should be: https://example.com/sse (replace with your actual Space URL)
• —No authentication is required for this public endpoint

3. Available MCP Tools

Once connected, ChatGPT will have access to:

• —get_documentation: Get up-to-date documentation for libraries and frameworks using Context7

4. Example Usage in ChatGPT

After connecting the MCP server, you can ask ChatGPT questions like:

• —"How do I create a FastAPI application?"
• —"Show me React hooks documentation"
• —"What's the latest Next.js documentation?"
• —"How do I use MongoDB with Python?"

ChatGPT will automatically use the Context7 MCP server to fetch current documentation and provide accurate, up-to-date information.

5. Troubleshooting

If the MCP connection fails:

1. 1.Ensure your Hugging Face Space is public and running
2. 2.Check that you're using the correct URL with /sse endpoint
3. 3.Verify the Space is not sleeping (visit the Space URL in browser first)
4. 4.Try refreshing ChatGPT Desktop and reconfiguring the MCP server