A centralized FastAPI server that provides a unified interface for all Model Context Protocol (MCP) tools. This hub allows you to easily manage and interact with multiple MCP servers through a single API endpoint.

• Centralized API: Single endpoint for all MCP tools
• Dynamic Server Management: Add/remove Docker MCP servers effortlessly
• Multiple Transport Support: stdio, http, sse
• Interactive Documentation: Swagger UI and ReDoc
• Health Monitoring: Built-in health checks and status endpoints
• Environment Variable Management: Automatic configuration from .env files
• Docker Support: Run MCP servers in containers with proper isolation

• Python 3.11+
• Docker (for MCP servers)
• Git

1. Clone the repository:

   git clone https://github.com/pranjal-pravesh/mcp-docker-hub.git
   cd mcp-docker-hub

2. Set up environment variables:

   cp env.example .env
   # Edit .env with your API keys

3. Install dependencies:

   python3 -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt
   pip install -e .

4. Start the hub:

   python -m mcp_hub.mcp_hub_server --host 0.0.0.0 --port 8000 --load-config

5. Access the API:

   • Main API: http://localhost:8000
   • Documentation: http://localhost:8000/docs
   • Health Check: http://localhost:8000/

1. Create a Google Cloud VM (f1-micro recommended for free tier):

   gcloud compute instances create mcp-hub \
     --zone=us-central1-a \
     --machine-type=f1-micro \
     --image-family=ubuntu-2204-lts \
     --image-project=ubuntu-os-cloud \
     --tags=http-server,https-server \
     --metadata=startup-script='#! /bin/bash
     curl -fsSL https://raw.githubusercontent.com/pranjal-pravesh/mcp-docker-hub/main/deploy-gcp.sh | bash'

2. SSH into the VM:

   gcloud compute ssh mcp-hub --zone=us-central1-a

3. Run the deployment script:

   curl -fsSL https://raw.githubusercontent.com/pranjal-pravesh/mcp-docker-hub/main/deploy-gcp.sh | bash

1. SSH into your VM:

   gcloud compute ssh YOUR_VM_NAME --zone=YOUR_ZONE

2. Clone and run the deployment script:

   git clone https://github.com/pranjal-pravesh/mcp-docker-hub.git
   cd mcp-docker-hub
   chmod +x deploy-gcp.sh
   ./deploy-gcp.sh

Allow port 8000 in your Google Cloud firewall:

gcloud compute firewall-rules create allow-mcp-hub \
  --allow tcp:8000 \
  --target-tags=http-server \
  --description="Allow MCP Hub traffic"

curl -X POST "http://localhost:8000/servers/add-docker" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "my-server",
    "docker_image": "mcp/my-server",
    "transport": "stdio",
    "env_vars": {"API_KEY": "your-key"}
  }'

python scripts/manage_mcp_servers.py add \
  --name my-server \
  --docker-image mcp/my-server \
  --transport stdio \
  --env-vars '{"API_KEY": "your-key"}'

Edit configs/mcp_servers.json:

{
  "my-server": {
    "docker_image": "mcp/my-server",
    "transport": "stdio",
    "env_vars": {
      "API_KEY": "your-key"
    }
  }
}

# Via API
curl -X DELETE "http://localhost:8000/servers/my-server"

# Via CLI
python scripts/manage_mcp_servers.py remove --name my-server

• GET / - Server status and health check
• GET /docs - Interactive API documentation
• GET /redoc - Alternative API documentation

• GET /servers - List all MCP servers
• POST /servers/start-all - Start all servers
• POST /servers/stop-all - Stop all servers
• POST /servers/add-docker - Add Docker MCP server
• DELETE /servers/{name} - Remove server

• GET /tools - List all available tools
• GET /tools/qwen - Get enabled tools in Qwen format (JSON lines)
• GET /tools/qwen?required_only=true - Get enabled tools with only required parameters
• POST /tools/call - Call any MCP tool
• GET /tools/info/{tool_name} - Get tool details

• GET /servers/configured - List configured servers
• POST /servers/load-config - Load from JSON config
• POST /servers/save-config - Save to JSON config
• GET /servers/check-availability - Check server availability

1. Docker Permission Denied:

   sudo usermod -aG docker $USER
   newgrp docker

2. Port Already in Use:

   sudo lsof -i :8000
   sudo kill -9 <PID>

3. Environment Variables Not Loading:

   python scripts/check_servers.py

4. Service Not Starting:

   sudo systemctl status mcp-hub.service
   sudo journalctl -u mcp-hub.service -f

If you're currently using Docker-in-Docker and experiencing permission issues:

# Run the migration script
./migrate-to-direct.sh

This will:

• Stop the Docker container
• Set up a Python virtual environment
• Create a systemd service
• Start the MCP Hub directly on the host

• Dynamic MCP Servers Guide
• Google Cloud Setup Guide
• Deployment Guide

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

This project is licensed under the MIT License - see the LICENSE file for details.