toltalchaos/graphql-MCP

Flask + MCP Server Setup

This project runs a Flask GraphQL server and an MCP (Model Context Protocol) server on different ports.

🐳 Docker Setup (Recommended) --desktop with MCP enabled

Quick Start with Docker Compose

# Build and run both servers in a container
docker-compose up --build

# Run in background
docker-compose up -d --build

# Stop the services
docker-compose down

Manual Docker Build

# Build the image
docker build -t flask-mcp-app .

# Run the container
docker run -p 5000:5000 -p 8001:8001 flask-mcp-app

🐍 Local Development Setup

Prerequisites

pip install -r requirements.txt

Option 1: Run Both Servers Together

python run_servers.py

This will start:

• Flask server on port 5000 (http://localhost:5000)
• MCP server on port 8001

Option 2: Run Servers Separately

Flask server only:

python run_servers.py flask
# or directly:
python app.py

MCP server only:

python run_servers.py mcp
# or directly:
python app.py mcp

🌐 Accessing the Services

• Flask GraphQL Server: http://localhost:5000

  • GraphQL endpoint: http://localhost:5000/graphql
  • Schema info: http://localhost:5000/schema_info

• MCP Server: localhost:8001

  • This is the MCP protocol server for AI tool integration

📝 GraphQL Usage

You can test the GraphQL endpoint with a query like:

query {
  resolveHello(name: "World") {
    greeting
  }
}

🛑 Stopping the Servers

Docker:

docker-compose down

Local Development:

• If running both servers together, press Ctrl+C to stop both
• If running separately, press Ctrl+C in each terminal

📁 Project Structure

├── app.py                 # Main application with Flask + MCP
├── run_servers.py         # Script to run both servers
├── requirements.txt       # Python dependencies
├── Dockerfile            # Docker configuration
├── docker-compose.yml    # Docker Compose configuration
├── .dockerignore         # Files to ignore in Docker build
└── README.md            # This file