MCP Demo

This project demonstrates Anthropic's Model Context Protocol (MCP) using Python and Docker. It showcases the client-server architecture of MCP with practical tool examples.

What is MCP?

The Model Context Protocol (MCP) is an open standard created by Anthropic for connecting AI models to external data sources and tools. It standardizes how applications provide context to Large Language Models (LLMs).

Project Structure

mcp-demo/
├── docker-compose.yml
├── .env.example        # Template for environment variables
├── .gitignore          # Git ignore file
├── CLAUDE.md           # Guidelines for Claude Code
├── server/
│   ├── Dockerfile
│   ├── requirements.txt
│   └── server.py
└── client/
    ├── Dockerfile
    ├── requirements.txt
    └── client.py

Components

1. MCP Server: Exposes tools and resources that can be used by an MCP client
2. MCP Client: Connects to the server and accesses its tools and resources

Features

The demo includes several tools:

1. greet - A simple greeting tool that returns a personalized message
2. add - A calculator that adds two numbers together
3. generate_story - A creative story generator with different themes and lengths

Prerequisites

• Docker and Docker Compose
• An Anthropic API key (optional, for advanced use cases)

Setup Instructions

1. Clone this repository:

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

2. (Optional) Create a .env file from the template:

   cp .env.example .env
   # Edit .env to add your Anthropic API key if needed
   

3. Start the containers with Docker Compose:

   docker-compose up --build
   

4. The client will connect to the server and demonstrate:

   • Listing available tools
   • Calling the greeting and addition tools
   • Generating creative stories with different themes and lengths

5. To stop the containers:

   docker-compose down
   

Extending the Demo

This demo can be extended in several ways:

1. Add more complex tools to the server (e.g., database access, external API integration)
2. Connect the client to an actual Claude model using the Anthropic API
3. Create a more interactive application with a web UI
4. Implement tool streaming for real-time updates

Troubleshooting

• Connection Issues: Ensure Docker networking is correctly set up. The client container should be able to reach the server on the Docker network.
• API Key: If using Anthropic API integration, ensure your API key is correctly set in the .env file.
• Logs: Check Docker logs for detailed error messages (docker-compose logs).

Resources

• Model Context Protocol Documentation
• Python MCP SDK
• Anthropic's Claude Documentation
• Docker Networking Guide