This project is an Agentic Retrieval-Augmented Generation (RAG) system that processes PDFs using Gemini as the model, LangChain for agent-based orchestration, and ChromaDB for vector storage and Streamlit for deploying an interactive web application. The project is built with GitHub Copilot assistance.
- Extracts relevant information from PDFs.
- Uses LangChain to create an intelligent agent.
- Stores and retrieves document embeddings with ChromaDB.
- Employs Gemini for natural language understanding and response generation.
- Provides an interactive web interface using Streamlit.
- Model: Gemini
- Framework: LangChain
- Database: ChromaDB
- Web Application: Streamlit
- Assistant: GitHub Copilot
- Add PDFs to the data/ folder.
- Run the script to process and query them.
- Customize the agent’s behavior in agentic-rag.py.
This project was inspired and guided by the following resource:
- An Improved Langchain RAG Tutorial (v2) by pixegami: This tutorial provided valuable insights into implementing a Retrieval-Augmented Generation system using LangChain and local LLMs.