Stackoverflow Teams MCP

A Model Context Protocol (MCP) server that provides access to the Stack Exchange API for Teams, allowing AI assistants to search and retrieve information from Stack Overflow.

Project Overview

This project exposes two main MCP tools:

1. stackoverflow_questions - Search Stack Overflow for questions and answers matching a query
2. stackoverflow_excerpts - Search Stack Overflow for excerpts matching a query

The server uses the Stack Exchange API to fetch data and formats the results in a structured way.

Project Structure

stackoverflow-teams-mcp/
├── .vscode/mcp.json # Dogfooding - Using this server with VSCode Copilot in this project
├── main.py          # MCP server logic
└── utils.py         # data models

Setup Instructions

This project uses uv for Python package management. Follow these steps to set up the project:

Prerequisites

• uv package manager

Installation

1. Clone the repository:

   git clone [your-repository-url]
   cd stackoverflow-teams-mcp
   

2. Create a virtual environment and install dependencies using uv:

   uv venv
   source .venv/bin/activate
   uv sync
   

Running the Server

For Development

Use fastmcp dev for local development.

It runs the inspector tool written by Anthropic.

fastmcp dev main.py

Supply these 2 env variables before starting the MCP server.

• BASE_URL (base url for the api)
• API_KEY (api key)

For using with VSCode Copilot

Check out .vscode/mcp.json