qs-first-mcp

qs-first-mcp

This repository provides a simplified implementation of Anthropic's Model Context Protocol (MCP) quickstart guide.

The repo achieves this goal by creating and centralizing the setup of a MCP tool, in this case, a basic weather server using Python, connected to Claude for Desktop.

System Requirements

• Python 3.10 or higher
• uv (download from Astral's repo)
• Claude for Desktop (download from Anthropic's website)

Setup

1. Clone and install dependencies

git clone git@github.com:SSGrady/qs-first-mcp.git
cd qs-first-mcp/weather
uv add "mcp[cli]" httpx

2. Configure Claude:

• On VSCode (or Cursor/Windsurf) open:

code ~/Library/Application\ Support/Claude/claude_desktop_config.json

• next, go ahead and add:

{
   "mcpServers": {
       "weather": {
           "command": "[UV_PATH]",
           "args": [
               "--directory",
               "~/qs-first-mcp/weather",
               "run",
               "weather.py"
           ]
       }
   }
}

💡 You can find UV_PATH by running which uv

3. Restart Claude for Desktop

💡 If you fix any bugs, make sure to repeat step 3

Usage

Try asking Claude about any USA weather:

• "What's the weather in Orlando?"
• "What are the active weather alerts in California?"

More Info

• This uses the National Weather Service API (USA only)
• See MCP debugging docs for help with any errors