🤖 Building AI Chat Agents with LangChain, Chainlit & GitHub Models

A 90-minute hands-on workshop for tech conferences

Welcome to this practical workshop where you'll build a fully functional AI chat agent from scratch! By the end of this session, you'll have created an intelligent assistant that can have conversations, call external tools (like fetching weather data), and integrate with the Model Context Protocol (MCP).

---

🎯 What You'll Build

A conversational AI agent with:

• Real-time chat interface powered by Chainlit
• LLM backbone using GitHub Models (completely free!)
• Tool calling capabilities (weather API integration)
• MCP integration for extensible agent capabilities

---

📚 Workshop Structure

Phase	Duration	Topic
0	5 min	Prerequisites & Setup Check
1	10 min	Environment Setup
2	10 min	GitHub Models Configuration
3	15 min	Your First Chainlit Chat App
4	15 min	Adding LangChain for LLM Power
5	20 min	Tool Calling with Weather API
6	15 min	MCP Integration

Each phase ends with a ✅ Checkpoint where you'll verify everything works before moving on.

---

🛠️ Tech Stack

Technology	Purpose	Version
Python	Runtime	3.10+
Chainlit	Chat UI Framework	2.x
LangChain	LLM Orchestration	0.3.x
GitHub Models	Free LLM API	Latest
WeatherAPI	Tool Demo	Free Tier
MCP	Agent Protocol	1.x

---

🚀 Quick Start

# Clone or create the workshop folder
mkdir langchain-chainlit-workshop
cd langchain-chainlit-workshop

# Follow the phases in order starting from:
# 📖 docs/00-prerequisites.md

---

📁 Project Structure

langchain-chainlit-workshop/
├── README.md                 # You are here
├── SUMMARY.md               # GitBook navigation
├── docs/
│   ├── 00-prerequisites.md  # What you need before starting
│   ├── 01-environment.md    # Python & dependencies setup
│   ├── 02-github-models.md  # Getting your free API access
│   ├── 03-chainlit-basics.md # First chat application
│   ├── 04-langchain.md      # LLM integration
│   ├── 05-tool-calling.md   # Weather API tools
│   └── 06-mcp-integration.md # MCP setup
├── solutions/               # Complete code for each phase
│   ├── phase-03/
│   ├── phase-04/
│   ├── phase-05/
│   └── phase-06/
└── assets/                  # Images and diagrams

---

👨‍🏫 For Workshop Facilitators

Before the Workshop

1. Ensure attendees have received the prerequisites checklist
2. Have backup GitHub tokens ready for attendees who forgot
3. Test the WeatherAPI endpoint is responding
4. Prepare a hotspot/backup internet option

During the Workshop

• Each phase has a checkpoint - don't proceed until 80%+ have passed
• The solutions/ folder contains working code for each phase
• Common issues are documented at the end of each phase

---

📝 License

This workshop material is provided under the MIT License. Feel free to use, modify, and distribute for your own workshops and training sessions.

---

Ready to start? Head to Prerequisites! 👉