This repository contains implementations and experiments with building AI-powered chatbots using LangGraph — a powerful framework for building stateful, multi-agent applications powered by language models.
The goal of this project is to create a modular, scalable chatbot architecture leveraging:
- 🧠 LangGraph for state management and agent flow
- 🛠️ Pydantic for data validation and structuring
- 🧰 Supporting tools and utilities to expand chatbot functionality
Modular and scalable AI chatbot architectures built using LangGraph and Pydantic.
This project demonstrates how to build multi-agent chatbots, define structured flows, and prototype intelligent workflows using LangGraph.
- ✅ Multi-agent chatbot workflows using LangGraph
- 🧠 State validation with Pydantic
- 🔄 Switchable architectures (e.g., Supervisor, Swarm)
- ⚙️ Custom tools and utilities
- 🧪 Easy to test via LangGraph Dev Studio
langgraph-chatbots/
├── demo\_supervisor/ # Supervisor architecture demo
│ ├── graph.py
│ └── state.py
├── demo\_swarm/ # Swarm-based agent demo
│ ├── graph.py
│ └── state.py
├── tools/ # Custom tools/utilities
│ └── custom\_tools.py
├── requirements.txt
├── .env # Your LLM credentials
└── README.md
git clone https://github.com/Archi-shaw/langgraph-chatbots.git
cd langgraph-chatbots
python -m venv venv
source venv/bin/activate
pip install -r requirements.txtCreate a .env file with your OpenAI credentials:
OPENAI_API_KEY=your-api-key-here
cd demo_supervisor
langgraph dev- Prototyping AI chatbot workflows
- Simulating multi-agent systems
- Exploring LangGraph capabilities
Have ideas for new architectures or improvements? PRs are welcome!