MCP Python Demo

This repository demonstrates the use of Model Context Protocol (MCP) SDK for Python.

Setup

If using Dev Container (Recommended)

If you're using the dev container (VS Code with Docker), everything is already set up for you! The container includes Python, uv, and all necessary tools.

If running locally

1. Install uv if you haven't already

2. Create a virtual environment:

   uv venv
   

3. Activate the virtual environment:

   • On macOS/Linux:

     source .venv/bin/activate
     

   • On Windows:

     .venv\Scripts\activate
     

4. Install dependencies from pyproject.toml:

   uv sync --dev
   

Running Development Inspector

To run the development inspector:

mcp dev server.py

Installing in Claude

To install this MCP plugin in Claude:

1. Run this command:

   mcp install server.py
   

2. Restart Claude

3. Troubleshooting: If you get an error, fix the uv path in the config to an absolute path:

   which uv
   

   Then update the configuration with the absolute path.

Installing in GitHub Copilot

📺 If you prefer learning from videos, watch this video from Burke Holland or this video from James Montemagno.

If you already installed in Claude Desktop

Enable chat.mcp.discovery.enabled: true in your settings and VS Code will discover existing MCP server lists, and proceed to use the tool in GitHub Copilot Agent mode.

If you did not install in Claude Desktop

• If you want to associate the MCP server only with a particular repo, create a .vscode/mcp.json file with this content:

  {
  "inputs": [
  ],
  "servers": {
        "pamelas-mcp": {
           "command": "PATH/TO/uv",
           "args": [
              "--directory",
              "/PATH/TO/mcp-python-demo",
              "run",
              "server.py"
           ]
        }
  }
  }
  

• Alternatively, if you want to associate the MCP server with all repos, add to your VS Code User Settings JSON:

  {
  "mcp": {
     "inputs": [],
     "servers": {
        "pamelas-mcp": {
        "command": "/PATH/TO/uv",
        "args": [
           "--directory",
           "/PATH/TO/mcp-python-demo",
           "run",
           "server.py"
        ]
        }
     }
  }
  }
  

  Another way to update settings is to run this command in the terminal:

  code-insiders --add-mcp "{\"name\":\"pamelas-mcp\",\"command\":\"/PATH/TO/uv\",\"args\":[\"--directory\",\"/PATH/TO/mcp-python-demo\",\"run\",\"server.py\"]}"
  

Using tools in Copilot

1. Now that the mcp server is discoverable, open GitHub Copilot and select the Agent mode (not Chat or Edits).
2. Select the "refresh" button in the top right corner of the Copilot chat text field to refresh the server list.
3. Select the "🛠️" button to see all the possible tools, including the ones from this repo.
4. Put a question in the chat that would naturally invoke one of the tools, for example: "What is the current weather in Seattle?".