AI Chatbot using LangChain, LangGraph, and OpenAI

This project is a terminal-based AI chatbot built with LangChain, LangGraph, and OpenAI. It uses the modern ReAct agent architecture and supports streaming responses directly to the terminal for a smooth and interactive experience.

Features

• Uses LangGraph's latest create_react_agent
• Supports real-time streaming responses
• Simple and clean terminal interface
• Easily expandable to include tools, memory, or additional models
• Uses the latest OpenAI gpt-4o-mini model
• Environment variables loaded using python-dotenv

Requirements

Install the required libraries:

pip install langchain langgraph langchain-openai python-dotenv

How to Use

1. Create an .env file in the project directory.
2. Add your OpenAI API key:

OPENAI_API_KEY=your_api_key_here

3. Run the script:

python main.py

4. Type messages in the terminal and the chatbot will respond.
5. Type exit or quit to close the program.

Project Structure

/project-folder
    main.py
    README.md
    .env

Code Explanation

• ChatOpenAI loads the OpenAI model for the chatbot.
• create_react_agent creates a modern ReAct agent using LangGraph.
• The agent streams output using agent.stream.
• HumanMessage is used for sending user input to the model.

Example Prompt Flow

You: Hello
Bot: Hi, how can I assist you?

Future Improvements

• Add memory support
• Add custom tools (search, calculator, file operations)
• Add voice input and text-to-speech output
• Create a full GUI or web interface