Skip to content

README.md

Latest commit

 

History

History
66 lines (40 loc) · 2.52 KB

README.md

File metadata and controls

66 lines (40 loc) · 2.52 KB
Raw

MCP Server - Weather

This is a sample MCP Server in Python implementing weather tools.

Get Started

Prerequisites

To run the MCP Server in your local dev machine, you will need:

Prepare Environment

Using uv

  • Create virtual environment

    uv venv

  • Run VSCode Command "Python: Select Interpreter" and select the python from created virtual environment

    (Note: reload VSCode and the terminal to ensure the virtual environment python is used)

  • Install dependencies (include dev dependencies)

    uv pip install -r pyproject.toml --extra dev

Or, using pip

  • Create virtual environment

    python -m venv .venv

  • Run VSCode Command "Python: Select Interpreter" and select the python from created virtual environment

    (Note: reload VSCode and the terminal to ensure the virtual environment python is used)

  • Install dependencies (include dev dependencies)

    pip install -e .[dev]

Debug in MCP Inspector

  • Open VSCode Debug panel. Select Debug in Inspector (Edge) or Debug in Inspector (Chrome). Press F5 to start debugging.
  • When MCP Inspector launches in the browser, click the Connect button to connect this MCP server.
  • Then you can List Tools, select a tool, input parameters, and Run Tool to debug you server code.
  • And ofcourse, you can add breakpoint to the tool implementation code.
More Details

When launching debugging, it runs two tasks

  • first, the MCP server is launched (by default on port 3001)
  • then, the MCP Inspector is launched (by default on port 5173 and 3000)

The whole definition can be found in tasks.json. You can also edit launch.json, tasks.json, __init__.py to change above ports.

In addition, the MCP Inspector is implemented in Node.js. To launch it within this Python project, there are two tricky steps:

  1. Put the bundled inspector client and server to inspector/. (Forked and added bundle logic to the inspector's source, see forked repo's branch)
  2. Launch VSCode executable as Node.js. See tasks.json and launcher.py.