Document-Based AI Chatbot (RAG)

Overview

This project is a document-based AI chatbot built using Flask, LangChain, FAISS, and open-source LLMs via Ollama.
It follows the Retrieval-Augmented Generation (RAG) architecture to answer queries from PDF documents.

Tech Stack

• Python
• Flask
• LangChain
• FAISS
• HuggingFace Embeddings
• Ollama (Mistral / Phi)

Workflow

1. PDFs are loaded and split into chunks
2. Embeddings are generated using sentence-transformers
3. Vectors are stored in FAISS
4. User queries retrieve relevant chunks
5. LLM generates grounded responses

How to Run

python ingest.py
python app.py