GitHub MCP Server Integration

GitHub MCP Server Integration

This project demonstrates a Model Context Protocol (MCP) Server that integrates with the GitHub API. It allows AI Assistants like Claude or OpenAI GPT to interact with GitHub using MCP.

---

🌐 Live Demo

👉 Deployed Server Link
👉 .well-known/ai-plugin.json endpoint:

https://github-mcp-server-production.up.railway.app/.well-known/ai-plugin.json

---

📖 Project Overview

This MCP Server exposes a REST API that allows AI Assistants to:

• Get user GitHub profile information
• List repositories for a user
• Create an issue in a repository

The server follows Model Context Protocol (MCP) standards to ensure interoperability with AI systems.

---

🔧 Tech Stack

• FastAPI (Python framework for building APIs)
• OpenAPI (Standard for describing REST APIs)
• ai-plugin.json (MCP metadata configuration)
• GitHub REST API v3
• Deployed on Railway

---

✨ Features

• MCP-compliant server with OpenAPI documentation
• Secure interaction with GitHub using a personal access token
• AI Assistants can:
  • Fetch GitHub user details
  • List repositories by username
  • Create issues on repositories

---

📂 Project Structure

├── ai-plugin.json             # MCP Plugin metadata for AI Assistants
├── openapi.yaml               # OpenAPI spec describing the available endpoints
├── main.py                    # FastAPI MCP server code
├── requirements.txt           # Python dependencies
└── README.md                  # Project documentation (this file)

---

🚀 How It Works

1. FastAPI Server

• Exposes REST API endpoints like:
  • /github/user
  • /github/repos/{username}
  • /github/repos/{owner}/{repo}/issues

2. ai-plugin.json

• Metadata for AI assistants to understand the MCP server:
  • Plugin name, description
  • Authentication method
  • OpenAPI URL reference

3. openapi.yaml

• Defines all routes and parameters for the AI assistant to interact with.

---

🔐 Authentication

• GitHub Personal Access Token (PAT)
  • Set as an environment variable: GITHUB_TOKEN
  • Use in your .env file or directly in Railway's environment settings.

---

⚙️ Setup Instructions (Local Development)

Prerequisites:

• Python 3.9+
• GitHub Personal Access Token (PAT) with repo permissions

Clone the repo:

git clone https://github.com/yourusername/github-mcp-server.git
cd github-mcp-server

Install dependencies:

pip install -r requirements.txt

Create .env file:

GITHUB_TOKEN=your_personal_access_token

Run the FastAPI server:

uvicorn main:app --reload

Visit: http://localhost:8000/docs for the Swagger UI

---

🌍 Deployment (Railway)

1. Login at Railway
2. Create a new project → Deploy from GitHub → Select your MCP repo
3. Add the GITHUB_TOKEN as an environment variable in Railway
4. Confirm the start command:

   uvicorn main:app --host 0.0.0.0 --port 8000
   

5. Deploy & get your production URL (e.g., https://github-mcp-server-production.up.railway.app)

---

🤖 Demonstration with AI Assistants

Claude AI / OpenAI GPT Plugin (optional based on access)

• Register your .well-known/ai-plugin.json URL in the AI assistant settings.
• Interact using natural language prompts:

  "List my GitHub repositories."
  "Create an issue in my repo named 'sample-repo'."
  

---

📜 OpenAPI Endpoints

Endpoint	Method	Description
/github/user	GET	Get authenticated user info
/github/repos/{username}	GET	List repositories of a user
/github/repos/{owner}/{repo}/issues	POST	Create an issue in a repo

---

📝 Example Requests

Get user info

curl -X GET https://github-mcp-server-production.up.railway.app/github/user

List repos

curl -X GET https://github-mcp-server-production.up.railway.app/github/repos/<username>

Create an issue

curl -X POST https://github-mcp-server-production.up.railway.app/github/repos/<owner>/<repo>/issues \
  -H "Content-Type: application/json" \
  -d '{"title": "Bug found!", "body": "Please fix this bug."}'

---