MCP 服务器

Weather MCP Server

Enables AI assistants to fetch real-time weather data, forecasts up to 10 days, and compare weather conditions between cities using WeatherAPI.com with support for both Celsius and Fahrenheit units.

README

Weather MCP Server 🌤️

A Model Context Protocol (MCP) server that provides weather information tools. This server enables AI assistants like Claude to fetch real-time weather data, forecasts, and compare weather across different cities.

Features

Current Weather: Get real-time weather information for any city
Weather Forecast: Get weather predictions for up to 10 days (based on API plan)
Weather Comparison: Compare weather conditions between two cities
Supports both Celsius (metric) and Fahrenheit (imperial) units

Prerequisites

Python 3.8 or higher
WeatherAPI.com API key (free tier available)

Installation

Clone the repository

git clone https://github.com/Abhishek3689/Weather_MCP_Server.git
cd weather-mcp-server

Install dependencies
```
pip install -r requirements.txt
```
Set up environment variables

Create a .env file in the project root:
```
WEATHER_API_KEY=your_api_key_here
```
Get your free API key from WeatherAPI.com

Usage

🚀 Running The Server (Important!) 👉 If you double-click or run the file, you will NOT see results.

MCP servers do not show output directly.

They wait for Claude to connect to them.

But you can still start the server manually to ensure it’s working:

python weather_server_mcp.py

You would see somthing like this it means server is running correclty

          ╭──────────────────────────────────────────────────────────────────────────────╮
          │                                                                              │
          │                         ▄▀▀ ▄▀█ █▀▀ ▀█▀ █▀▄▀█ █▀▀ █▀█                        │
          │                         █▀  █▀█ ▄▄█  █  █ ▀ █ █▄▄ █▀▀                        │
          │                                                                              │
          │                                FastMCP 2.13.1                                │
          │                                                                              │
          │                                                                              │
          │                    🖥  Server name: Weather Server                            │
          │                                                                              │
          │                    📦 Transport:   STDIO                                     │
          │                                                                              │
          │                    📚 Docs:        https://gofastmcp.com                     │
          │                    🚀 Hosting:     https://fastmcp.cloud                     │
          │                                                                              │
          ╰──────────────────────────────────────────────────────────────────────────────╯

Integration with Claude Desktop

To use this server with Claude Desktop, add the following to your Claude Desktop configuration:

On macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

On Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "weather": {
      "command": "python",
      "args": [
        "weather_server_mcp.py"
      ],
      "env": {
        "WEATHER_API_KEY": "your_api_key_here"
      }
    }
  }
}

Note : for 'command' write absolute location instead of only "python" by checking using in command prompt

where python

you will see somthing like this

C:\Users\YourName\AppData\Local\Programs\Python\Python312\python.exe

After updating the config, restart Claude Desktop.

Available Tools

1. Get Current Weather

get_current_weather(city: str, units: str = "metric")

city: City name (e.g., 'London', 'New York', 'Tokyo')
units: 'metric' for Celsius or 'imperial' for Fahrenheit

Example:

get_current_weather("London", "metric")

2. Get Weather Forecast

get_forecast(city: str, days: int = 3, units: str = "metric")

city: City name
days: Number of forecast days (1-10)
units: Temperature units

Example:

get_forecast("Paris", days=5, units="metric")

3. Compare Weather

compare_weather(city1: str, city2: str, units: str = "metric")

city1: First city name
city2: Second city name
units: Temperature units

Example:

compare_weather("Tokyo", "New York", "metric")

Project Structure

weather-mcp-server/
├── weather_server_mcp.py   # Main MCP server implementation
├── requirements.txt        # Python dependencies
├── .env                    # Environment variables (not committed)
├── .gitignore             # Git ignore file
└── README.md              # This file

API Information

This project uses the WeatherAPI.com service:

Free tier: 1,000,000 calls/month
3-day forecast on free plan
Real-time weather data
Historical weather data

Example Output

Current Weather in London, United Kingdom

📍 Location: 51.52, -0.11
🌡️  Temperature: 15°C
🤔 Feels Like: 14°C
☁️  Condition: Partly cloudy
💧 Humidity: 72%
🔽 Pressure: 1013.0 hPa
💨 Wind Speed: 15.1 kph
☁️  Cloudiness: 25%
👁️  Visibility: 10.0 km