A hands-on workshop teaching how to build AI chat agents using Microsoft Agent Framework, Chainlit, and GitHub Models.
A fully functional AI chat agent with:
- Web-based chat interface (Chainlit)
- Conversation memory
- Tool calling (weather API)
- MCP (Model Context Protocol) integration
- Python 3.10+
- GitHub account (for GitHub Models access)
- Basic Python knowledge
| Phase | Topic | Time |
|---|---|---|
| 1 | Environment Setup | 10 min |
| 2 | GitHub Models Connection | 10 min |
| 3 | Chainlit Chat Interface | 15 min |
| 4 | Microsoft Agent Framework Agents | 15 min |
| 5 | Tool Calling | 20 min |
| 6 | MCP Integration | 15 min |
Total time: ~85 minutes
# Clone the repo
git clone <repo-url>
cd agent-framework-chainlit-workshop
# Create virtual environment
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
# Install dependencies
pip install -r requirements.txt
# Set up environment variables
cp .env.example .env
# Edit .env with your GITHUB_TOKEN
# Run the final solution
chainlit run solutions/phase-06/app.py -wQuick links:
- Start: Phase 1: Environment Setup
- Help: Troubleshooting Guide
- Microsoft Agent Framework - Microsoft's framework for building AI agents
- Chainlit - Build conversational AI interfaces
- GitHub Models - Free access to AI models
MIT