README.md

metadata

title: Agentic AI-powered chatbot with tools
emoji: 🐨
colorFrom: blue
colorTo: red
sdk: streamlit
sdk_version: 1.42.0
app_file: app.py
pinned: false
license: apache-2.0
short_description: Agentic AI chatbot using LangGraph and LangChain tools.

Agentic Chatbot with LangGraph, LangChain, and Python

This project demonstrates how to build an Agentic AI-powered chatbot with tools using LangGraph, LangChain, and Python. The chatbot leverages agentic workflows to handle complex, multi-step conversations with users.

Features

• Agentic Reasoning: The chatbot can plan, reason, and execute tasks using agent-based logic.
• Modular Design: Built with LangGraph for flexible conversational flows.
• Extensible: Easily add new tools, memory, or integrations.
• Supports OpenAI and Groq Models: Works with both OpenAI and Groq LLMs via API keys.

Requirements

• Python 3.8+
• langchain
• langgraph
• streamlit
• (Optional) OpenAI API key or Groq API key

Usage

1. Clone the repository:

   git clone https://github.com/abhinavranjan-ai/agenticai-chatbot-with-tools.git
   

Installation

pip install -r requirements.txt

2. Configure your environment:

   • Launch the app and enter your OpenAI or Groq API key in the Streamlit UI when prompted.

3. Run the chatbot app:

   streamlit run app.py
   

How It Works

• LangChain provides the language model interface and agent tools.
• LangGraph manages the conversational flow as a graph, enabling complex agentic behaviors.
• Streamlit powers the user interface, allowing users to input their API keys and interact with the chatbot.
• The chatbot receives user input, reasons about the task, and responds or takes actions accordingly.

Example

User: What's the weather in Paris and set a reminder for tomorrow.
Bot: The weather in Paris is sunny. Reminder set for tomorrow.

Customization

• Add new tools or memory modules in chatbot.py.
• Modify the agent's reasoning logic using LangChain's agent framework.

Try It Online

You can try the chatbot app instantly in your browser via Hugging Face Spaces.

Note: If you find this project helpful, feel free to follow me on LinkedIn!