Weather MCP Server

A Model Context Protocol (MCP) server that provides weather information for various cities.

Overview

This is a TypeScript-based MCP server that implements weather-related tools and resources. It enables AI agents and tools compatible with the MCP protocol to access weather data in a standardized way.

Features

• get_weather tool: Retrieve current weather conditions for supported cities
• weather://cities resource: List of supported cities with available weather data

Supported Cities

• London (22° C, Sunny)
• Paris (25° C, Cloudy)

Prerequisites

• Node.js 18+
• npm or yarn

Installation

# Install dependencies
npm install

Development

Build

# Compile TypeScript to JavaScript
npx tsc

Run

# Execute the server
node server.js

Or use tsx for direct execution:

npx tsx server.ts

Project Structure

.
├── server.ts              # Main server implementation with tools and resources
├── package.json           # Project dependencies and metadata
├── package-lock.json      # Locked dependency versions
├── .gitignore            # Git ignore rules
└── .github/
    └── copilot-instructions.md  # AI agent development guidelines

Architecture

This project follows the MCP server architecture:

• Server Setup: Uses McpServer from the official MCP SDK
• Tool Registration: Weather tools defined with Zod input validation
• Resource Management: Static resources listing available weather data
• Type Safety: Full TypeScript with strict mode

API Reference

Tools

get_weather

Retrieves current weather for a specified city.

Input Schema:

{
  city: string  // City name (e.g., "London", "Paris")
}

Response:

{
  "temp": "22° C",
  "conditions": "Sunny"
}

Resources

weather://cities

Returns a list of supported cities for weather queries.

Configuration

All configuration is in package.json. The server uses:

• @modelcontextprotocol/sdk - MCP protocol implementation
• zod - Runtime schema validation
• typescript - Type safety and compilation

Development Guidelines

See .github/copilot-instructions.md for detailed instructions for AI agents and developers working on this project.

Contributing

When adding new tools:

1. Define the input schema with Zod
2. Register the handler with proper MCP ContentBlock structure
3. Update this README with new features

License

ISC