A stateful agentic AI chatbot built using LangGraph, demonstrating how to design, route, and manage multi-agent workflows using graph-based execution.
This project focuses on agent orchestration, state management, and controlled AI behavior, rather than a simple single-prompt chatbot.
AgenticChatbot showcases how to build an agentic AI system where:
- Agents are nodes in a graph
- State flows across nodes
- Decisions determine the next execution path
Using LangGraph, the chatbot maintains context, handles branching logic, and executes actions in a controlled, explainable manner.
Traditional LLM chains are linear.
LangGraph enables:
- 🔄 Stateful execution
- 🧭 Conditional routing
- 🧠 Multi-agent coordination
- 🛑 Controlled autonomy
- 📊 Graph-based reasoning flow
This makes it ideal for agentic systems, AI workflows, and production-grade AI agents.
Each node reads and writes to a shared state object, enabling memory, control, and traceability.
- 🧠 Stateful AI Agents using LangGraph
- 🔀 Graph-based decision routing
- 🧩 Modular agent nodes
- 🛠️ Tool/reasoning separation
- 📈 Scalable agent workflows
- 🧪 Easy experimentation with agent logic
| Technology | Purpose |
|---|---|
| Python | Core language |
| LangGraph | Agentic graph execution |
| LangChain | LLM abstraction |
| OpenAI / LLM API / Groq | Reasoning engine |
| dotenv | Environment management |
git clone https://github.com/Nandinibajaj28/AgenticChatbot.git
cd AgenticChatbot
Install dependencies:
```bash
pip install -r requirements.txt