Weather MCP Example

An example Weather MCP repository illustrating the fundamentals of the Model Context Protocol, including tool definitions, server setup, and querying weather data through an MCP interface.

Features

• Fetch weather data for a given city
• Use the Open Meteo API and Open Geocoding API to get the weather data
• Use the standard input/output (stdio) transport for local development

How It Works

The MCP server is defined in the index.ts file usinf McpServer class from the MCP SDK. It registers a tool to fetch the weather of a city using the geocoding API to get the latitude and longitude of the city and then uses the Open Meteo API to get the weather data.

Then a transport is created to communicate with the server using the standard input/output (stdio) transport for local development and the HTTP transport for remote deployment. That information will be used by the AI agent to send better responses to the user.

How to Test

MCP Inspector

Using https://modelcontextprotocol.io/docs/tools/inspector#mcp-inspector you can test the MCP server locally.

Run npx -y @modelcontextprotocol/inspector npx -y tsx index.ts to start the MCP server and then you can test the tool using the MCP inspector.

npx -y @modelcontextprotocol/inspector to start the MCP inspector and npx -y tsx index.ts is how to run your file.

Claude Code

• Go to the Claude Code settings
• Developers
• Edit Config
  • Open the config file "claude_desktop_config.json"

{
  "mcpServers": {
    "weatherMCP": {
      "command": "npx",
      "args": [
        "-y",
        "tsx",
        "<path to your MCP server file>"
      ]
    }
  }
}

Cursor

• Go to the Cursor settings
• Go to the "Tools & MCP" section
• Click "Add Custom MCP" button
• Paste the following in the mcp.json file related to your MCP server:

{
  "mcpServers": {
    "weatherMCP": {
      "command": "npx",
      "args": [
        "-y",
        "tsx",
        "<path to your MCP server file>"
      ]
    }
  }
}