MCP 101: Building Context-Aware AI Agents

A Quarto presentation teaching engineers how to build applications using LLMs with the Model Context Protocol (MCP).

Quick Start

Prerequisites

• Quarto installed
• Python 3.10+ with uv package manager

View the Presentation

quarto preview

Run the Demo

1. Set up environment variables:

   # Copy the example environment file
   cp demo/.env.example demo/.env

2. Configure API keys in demo/.env:

   • OPENAI_API_KEY: Your OpenAI API key (get from OpenAI Dashboard)
   • GITHUB_PERSONAL_ACCESS_TOKEN: GitHub token with gist creation permissions (see setup below)

3. Install dependencies and run:

   # Install dependencies
   uv sync
   
   # Start the MCP server
   uv run demo/server.py
   
   # In another terminal, run the client
   uv run demo/client.py

GitHub Personal Access Token Setup

1. Go to GitHub Settings > Developer settings > Personal access tokens
2. Click "Generate new token" → "Generate new token (classic)"
3. Give it a descriptive name (e.g., "MCP 101 Demo")
4. Select the gist scope (required for creating gists)
5. Click "Generate token"
6. Copy the token and add it to your demo/.env file

What You'll Learn

• Bridge the gap between LLMs and real-world data
• Build MCP servers with tools, resources, and prompts
• Implement authentication and progress tracking
• Navigate the growing MCP ecosystem

Demo Features

The demo shows ArXiv paper search integrated with GitHub gist creation through MCP.