Modular MCP Server

This is a modular, auto-discovering Model Context Protocol (MCP) server built with Python and FastAPI. It's designed to be a scalable foundation for providing tools to Large Language Models (LLMs).

Features

• Auto-Discovery: Automatically loads tools from the /tools directory on startup.
• Modular: Tools are self-contained in their own files.
• Configurable: Uses a .env file for all configuration.
• Secure: Includes best practices like read-only database users and whitelisting.
• Performant: Runs on a high-performance Unix socket and uses an async database connection pool.

Setup

1. Clone the repository:

   git clone <repository_url>
   cd mcp
   

2. Create and activate a virtual environment:

   python3 -m venv .venv
   source .venv/bin/activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Configure your environment:

   cp .env.example .env
   

   Now, edit the .env file with your actual database credentials and settings.

Running the Server

To run the server on a Unix socket:

uvicorn mcp_server:app --uds /tmp/mcp.sock

Testing

You can test the running server using curl.

Discover tools:

curl --unix-socket /tmp/mcp.sock http://localhost/mcp/v1/tools | jq

Execute a tool:

curl -X POST \
  --unix-socket /tmp/mcp.sock \
  -H "Content-Type: application/json" \
  -d '{"arguments": {"timezone": "Europe/Istanbul"}}' \
  http://localhost/mcp/v1/tools/get_current_time:execute | jq