Agentic Chatbot

A production-ready scaffold for an agentic chatbot powered by LangChain and LangGraph, featuring FAISS vector search for document retrieval, Tavily web search integration for real-time information, and a Streamlit-based chat interface — all in a modular, extensible architecture.

Features

• Agentic reasoning — ReAct agent pattern for tool selection and execution
• Vector search — FAISS-powered document retrieval from local knowledge bases
• Web search — Tavily API integration for real-time internet queries
• Streaming responses — real-time token-by-token LLM output
• Modular architecture — pluggable tools and configurable agent behaviors
• Streamlit UI — clean, responsive chat interface
• Session management — conversation history and context persistence

Tech Stack

• Python 3 — Core programming language
• LangChain — LLM integration and tool-calling framework
• LangGraph — Stateful multi-agent orchestration
• FAISS — Vector similarity search
• Tavily — Web search API
• Streamlit — Web UI framework

Getting Started

git clone https://github.com/giovanniromero-dev/agentic-chatbot.git
cd agentic-chatbot
pip install -r requirements.txt
# Configure your .env with API keys
streamlit run app.py

---

Built with dedication by Giovanni Romero