What is MCP?
Model Context Protocol — Complete Guide 2026

# Step 1: Install the MCP Python library
# Run in terminal: pip install mcp

from mcp.server import Server, NotificationOptions
from mcp.server.models import InitializationOptions
import mcp.server.stdio
import mcp.types as types
import asyncio, httpx

# Create the server (give it a name)
server = Server("weather-server")

# Define a TOOL that Claude can call
@server.list_tools()
async def handle_list_tools():
    return [
        types.Tool(
            name="get_weather",
            description="Get current weather for any city",
            inputSchema={
                "type": "object",
                "properties": {
                    "city": {"type": "string", "description": "City name"}
                },
                "required": ["city"]
            }
        )
    ]

# Handle when Claude actually calls the tool
@server.call_tool()
async def handle_call_tool(name: str, arguments: dict):
    if name == "get_weather":
        city = arguments["city"]
        # Call real weather API
        async with httpx.AsyncClient() as client:
            resp = await client.get(
                f"https://wttr.in/{city}?format=3"
            )
        return [types.TextContent(type="text", text=resp.text)]

# Run the server (starts listening for Claude)
async def main():
    async with mcp.server.stdio.stdio_server() as (r, w):
        await server.run(r, w, InitializationOptions(
            server_name="weather-server",
            server_version="0.1.0"
        ))

asyncio.run(main())

{
  "mcpServers": {
    "weather": {
      "command": "python",
      "args": ["/path/to/your/weather_server.py"]
    }
  }
}