Model Context Protocol (MCP) Example

🎯 MCP Examples - Enterprise Edition

🚀 FastMCP 2.0 powered examples for building sophisticated Model Context Protocol servers

Welcome to the ultimate MCP examples repository! Build enterprise-grade AI servers with our living Minesweeper example and comprehensive FastMCP 2.0 templates.

✨ What's New in This Edition

🎮 Living Game Example: Watch LLMs master Minesweeper through logical reasoning
🚀 FastMCP 2.0: Enterprise features with multiple transport protocols
🏢 Production Ready: Authentication, sessions, real-time data, WebSocket support
🤖 AI Agent Friendly: Complete templates and guides for automated server generation

🎯 Featured: Minesweeper MCP Server

The ultimate demonstration of AI logical reasoning!

# Start the Minesweeper server
mcp dev examples/servers/minesweeper_server.py

# Or run as HTTP server
python examples/servers/minesweeper_server.py --transport http --port 8000

Watch an LLM play Minesweeper:

# Start a new expert game
new_game("expert")  # 30x16 grid, 99 mines!

# Get AI-powered strategic hints  
get_hint(game_id)   # Returns probability analysis

# Make strategic moves
reveal(game_id, x, y)  # Uncover cells
flag(game_id, x, y)    # Mark mines

# Analyze the board
analyze_board(game_id)  # Full probability matrix

🚀 Quick Start

Prerequisites

• Python 3.10+ (FastMCP 2.0 requirement)
• Poetry (recommended) or pip

Installation

# Clone the repository
git clone https://github.com/yourusername/mcp-examples.git
cd mcp-examples

# Install with Poetry (recommended)
poetry install

# Or with pip
pip install -e .

🎮 Try the Minesweeper Server

# Development mode with MCP inspector
mcp dev examples/servers/minesweeper_server.py

# HTTP server for web integration
python examples/servers/minesweeper_server.py --transport http --port 8000

# Test with the basic client
python examples/clients/basic_client.py

📁 Project Structure

examples/
├── servers/
│   ├── template_server.py       # 📄 Clean template - your starting canvas!
│   ├── minesweeper_server.py    # 🎯 Complex example - learn from this!
│   ├── basic_server.py          # Simple MCP server (official SDK)
│   ├── filesystem_server.py     # File operations server
│   └── comprehensive_server.py  # Full feature showcase
├── clients/
│   └── basic_client.py          # MCP client examples
└── configs/
    ├── claude_desktop_config.json
    └── README.md

# Documentation
├── FASTMCP_AI_AGENT_GUIDE.md    # Complete FastMCP 2.0 guide
├── AI_AGENT_INSTRUCTIONS.md     # Instructions for AI agents
└── CLAUDE.md                    # Project guidelines

🎮 Minesweeper Server Features

🧠 AI-Powered Gameplay

• Strategic Tools: new_game(), reveal(), flag(), get_hint()
• Analysis Tools: analyze_board(), probability calculations
• Learning Resources: Strategy guides, pattern libraries, tutorials

🏢 Enterprise Features

• Multiple Transports: stdio, HTTP, WebSocket ready
• Session Management: Multiple concurrent games
• Real-time Updates: Live game state streaming
• Authentication Ready: OAuth 2.0 support built-in
• Statistics: Global leaderboards and analytics

📚 Educational Resources

• Strategy Guide: Master-level Minesweeper techniques
• Pattern Library: Common configurations and solutions
• AI Prompts: Teaching assistant for learning players
• Probability Analysis: Mathematical approach to decision making

🔧 Server Examples

🎯 Minesweeper (FastMCP 2.0)

from fastmcp import FastMCP

mcp = FastMCP("🎯 Minesweeper Pro")

@mcp.tool
def new_game(difficulty: str = "beginner") -> dict:
    """🎮 Start a new Minesweeper game!"""
    game_id = game_engine.create_game(difficulty)
    return {"game_id": game_id, "message": "Good luck! 🎯"}

@mcp.resource("game://state/{game_id}")
def get_game_state(game_id: str) -> str:
    """Real-time game state resource."""
    return format_game_display(game_id)

@mcp.prompt
def strategy_guide(situation: str) -> str:
    """Generate strategic analysis for current situation."""
    return f"Strategic guidance for: {situation}..."

🔧 Basic Server (Official SDK)

from mcp.server.fastmcp import FastMCP

mcp = FastMCP("Basic Example Server")

@mcp.tool()
def add_numbers(a: float, b: float) -> float:
    """Add two numbers together."""
    return a + b

@mcp.resource("info://server")
def get_server_info() -> str:
    """Server information resource."""
    return "Basic MCP Server Example..."

🤖 For AI Agents

This repository is optimized for AI agents to build MCP servers!

🎯 Perfect Learning Setup

• 📄 template_server.py - Clean starting canvas with all boilerplate
• 🎮 minesweeper_server.py - Complex working example to learn from
• 📚 Complete Guides - Step-by-step instructions and patterns

🚀 Quick Start for AI Agents

# Copy the template and start building
cp examples/servers/template_server.py my_awesome_server.py

# Study the complex example for patterns
cat examples/servers/minesweeper_server.py

# Read the building guides
cat FASTMCP_AI_AGENT_GUIDE.md

📖 Complete Documentation

• FastMCP 2.0 AI Agent Guide - Enterprise server building
• AI Agent Instructions - Step-by-step templates
• Template Server - Ready-to-use starting point

🎯 Use Cases

• Game Servers: Interactive entertainment with real-time state
• Data Processing: Complex analysis with streaming results
• API Integration: External service coordination
• Educational Tools: Interactive learning systems
• Productivity Apps: Task management and automation

🏢 Enterprise Features (FastMCP 2.0)

🚀 Multiple Transports

# Development: stdio (default)
mcp.run()

# Production: HTTP with streaming
mcp.run(transport="streamable-http", host="0.0.0.0", port=8000)

# WebSocket support ready
mcp.run(transport="websocket", host="127.0.0.1", port=8080)

🔐 Authentication & Security

@mcp.set_auth_config
def auth_config():
    return {
        "authorization_url": "https://your-auth.com/oauth/authorize",
        "token_url": "https://your-auth.com/oauth/token",
        "client_id": os.getenv("CLIENT_ID"),
        "scopes": ["read", "write", "admin"]
    }

📊 Real-time Data & Sessions

@mcp.tool
async def create_session(user_id: str, data: dict) -> dict:
    """Create stateful user sessions."""
    session_id = str(uuid.uuid4())
    _sessions[session_id] = {"user": user_id, "data": data}
    return {"session_id": session_id}

@mcp.resource("stream://live-data")
async def live_data() -> str:
    """Streaming real-time data."""
    return f"Live update: {time.time()}"

🔧 Development Commands

# Development & Testing
mcp dev examples/servers/minesweeper_server.py
mcp dev examples/servers/basic_server.py

# HTTP Servers
python examples/servers/minesweeper_server.py --transport http --port 8000
python examples/servers/comprehensive_server.py --http --port 8080

# Install for Claude Desktop
mcp install examples/servers/minesweeper_server.py

# Run tests
poetry run pytest
poetry run pytest --cov=examples

# Code quality
poetry run black .
poetry run isort .
poetry run ruff check .
poetry run mypy .

🎯 Claude Desktop Integration

1. Copy configuration from examples/configs/claude_desktop_config.json
2. Update server paths to match your setup:

{
  "mcpServers": {
    "minesweeper": {
      "command": "python",
      "args": ["/path/to/examples/servers/minesweeper_server.py"]
    },
    "filesystem": {
      "command": "python", 
      "args": ["/path/to/examples/servers/filesystem_server.py"]
    }
  }
}

3. Restart Claude Desktop and start playing! 🎮

🎮 Minesweeper Gameplay Examples

🎯 Start Playing

# Create a new game
result = new_game("intermediate")  # 16x16, 40 mines
game_id = result["game_id"]

# Make your first move
reveal(game_id, 8, 8)  # Start in the center

# Get strategic advice
hint = get_hint(game_id)
# Returns: {"action": "reveal", "x": 5, "y": 3, "reason": "0% mine probability"}

🧮 Advanced Analysis

# Full probability analysis
analysis = analyze_board(game_id)
# Returns probability matrix for all hidden cells

# Strategic guidance
prompt = analyze_strategy(game_id, "stuck on complex constraint")
# Returns step-by-step strategic guidance

📊 Statistics & Learning

# Check your progress
stats = list_games()
# Shows win rate, best times, active games

# Learn from the masters
guide = get_strategy_guide()
# Comprehensive strategy documentation

🧪 Testing Your Servers

🔍 MCP Inspector

# Test any server interactively
mcp dev examples/servers/minesweeper_server.py
mcp dev examples/servers/basic_server.py

🤖 Programmatic Testing

# Test with FastMCP's built-in testing
from fastmcp.testing import create_test_client

async def test_minesweeper():
    client = create_test_client(mcp)
    
    # Test game creation
    result = await client.call_tool("new_game", {"difficulty": "beginner"})
    assert result["success"] == True
    
    # Test resources
    info = await client.get_resource("stats://global")
    assert "Statistics" in info

🌟 What Makes This Special

🎮 Living Examples

• Interactive Gameplay: Watch LLMs learn and strategize
• Real-time State: See decision-making processes unfold
• Educational Value: Understand AI reasoning through play

🏢 Enterprise Ready

• Production Deployment: HTTP servers, authentication, sessions
• Scalable Architecture: Multiple transports, WebSocket support
• Monitoring & Analytics: Built-in statistics and performance tracking

🤖 AI Agent Optimized

• Complete Templates: Copy-paste server generators
• Comprehensive Docs: Everything needed for autonomous building
• Best Practices: Security, error handling, testing patterns

📚 Learn More

• FastMCP Documentation - Official FastMCP 2.0 docs
• MCP Specification - Protocol specification
• Official MCP SDK - Python SDK
• MCP Community - Community resources

🤝 Contributing

We welcome contributions! Whether it's:

• 🎮 New game examples (Chess, Tic-tac-toe, Sudoku)
• 🏢 Enterprise patterns (Authentication, monitoring, deployment)
• 🤖 AI agent tools (Templates, generators, testing)
• 📚 Documentation (Tutorials, guides, examples)

Guidelines

• Keep examples practical and educational
• Include comprehensive documentation
• Follow security best practices
• Add proper error handling
• Use type hints throughout

🏆 Showcase

Built something cool with these examples? We'd love to feature it!

• Enterprise MCP deployments
• Educational AI systems
• Interactive game servers
• Creative LLM applications

📄 License

MIT License - build whatever you want! See LICENSE for details.

🙏 Acknowledgments

• Anthropic - For creating MCP and making AI better
• Jeremiah Lowin - For the amazing FastMCP framework
• MCP Community - For the excellent ecosystem
• Contributors - Who make this repository awesome

---

🎯 Ready to Build?

# 📄 Start with the clean template (recommended)
cp examples/servers/template_server.py my_server.py
mcp dev my_server.py

# 🎮 Study the complex example
mcp dev examples/servers/minesweeper_server.py

# 📚 Read the building guides
cat FASTMCP_AI_AGENT_GUIDE.md

# 🚀 Test your creation
python my_server.py --transport http --port 8000

🚀 Happy building! Let's make MCP servers that are both powerful and fun! 🎮