MCP (Model Context Protocol)

This repository starts with an implementation of the tutorial MCP servers and then extends it with custom examples. The repository is an exploration of additional supporting capabilities required to productionize remote MCP servers.

Overview

This project consists of multiple server implementations that provide tools for interacting with different services:

• Markdownify Server: tutorial server for markdown processing and conversion utilities
• Weather Server: tutorial server for weather data and forecasting services
• GitHub Server: implementation of the official GitHub MCP server
• AWS Resource Explorer: custom server for exploring and managing AWS resources
• Strava Server: custom server with tools for accessing and analyzing Strava activities

Project Structure

.
├── client-mcp/
├── server-aws-resource-explorer/
├── server-github/
├── server-markdownify/
├── server-strava/
├── server-weather/
├── ops/
└── requirements.txt

Setup

1. Install dependencies:

uv pip install --system -r requirements.txt

2. Configure environment variables for each service as needed (see individual server READMEs for details)

3. Run servers using the uv command:

uv run --with mcp[cli] mcp run <server-path>

Refer to the README files in each server directory for additional details.

Additional Resources

• Building an MCP server as an API developer - blog post that details the process for building out server-strava.
• Authoring agentic applications: to code or to prompt? - follow-up blog post that works through analysis and reporting for Strava data.