MCP (Model Context Protocol) Project

This project demonstrates the Model Context Protocol (MCP) implementation with a client-server architecture. It includes a weather service MCP server and a client that can interact with it using Anthropic's Claude AI model.

Overview

The Model Context Protocol (MCP) is a standard for AI assistants to communicate with external data sources and tools. This project showcases:

• MCP Weather Server: A .NET-based MCP server that provides weather data tools
• MCP Client: A .NET client that connects to MCP servers and integrates with Anthropic's Claude AI

Project Structure

MCP Server/
├── QuickstartClient/          # MCP Client application
│   ├── Program.cs            # Main client application
│   └── QuickstartClient.csproj
└── weather/                  # MCP Weather Server
    ├── Program.cs            # Server entry point
    ├── WeatherTools.cs       # Weather service tools
    └── weather.csproj

Features

Weather Server Tools

The weather server provides two main tools:

1. GetAlerts: Retrieves active weather alerts for a US state

   • Parameter: state (string) - The US state to get alerts for
   • Returns: Formatted weather alerts with event type, area, severity, and instructions

2. GetForecast: Gets weather forecast for a specific location

   • Parameters:
     • latitude (double) - Latitude of the location
     • longitude (double) - Longitude of the location
   • Returns: Detailed weather forecast with temperature, wind, and conditions

MCP Client

The client application:

• Connects to MCP servers via stdio transport
• Integrates with Anthropic's Claude AI model
• Provides an interactive chat interface
• Supports multiple server types (Python, Node.js, .NET)

Prerequisites

• .NET 9.0 SDK
• Anthropic API key (for the client)

Installation

1. Clone the repository

   git clone <repository-url>
   cd "MCP Server"
   

2. Set up the Weather Server

   cd weather
   dotnet restore
   dotnet build
   

3. Set up the Client

   cd ../QuickstartClient
   dotnet restore
   dotnet build
   

4. Configure API Keys

   For the client, you need to set up your Anthropic API key. You can do this in several ways:

   Option 1: User Secrets (Recommended)

   cd QuickstartClient
   dotnet user-secrets set "ANTHROPIC_API_KEY" "your-api-key-here"
   

   Option 2: Environment Variable

   set ANTHROPIC_API_KEY=your-api-key-here
   

Usage

Running the Weather Server

1. Start the weather server:

   cd weather
   dotnet run
   

   The server will start and listen for connections via stdio transport.

Running the Client

1. Start the client with the weather server:

   cd QuickstartClient
   dotnet run ../weather/weather.csproj
   

2. Interactive Chat Interface

   Once started, you'll see:

   Connected to server with tools: GetAlerts
   Connected to server with tools: GetForecast
   MCP Client Started!
   Enter a command (or 'exit' to quit):
   >
   

3. Example Commands

   You can now interact with the weather tools through natural language:

   • "Get weather alerts for California"
   • "What's the forecast for New York City? Use coordinates 40.7128, -74.0060"
   • "Show me weather alerts for Texas"

Running with Other Server Types

The client supports different server types:

• Python servers: dotnet run script.py
• Node.js servers: dotnet run script.js
• .NET projects: dotnet run project.csproj

Architecture

MCP Server (Weather)

The weather server implements the MCP protocol using:

• Transport: StdioServerTransport for communication
• Tools: WeatherTools class with annotated methods
• HTTP Client: Configured to access weather.gov API

Key components:

• Program.cs: Server configuration and startup
• WeatherTools.cs: Tool implementatio

... View full README on GitHub