RAG Document Intelligence System

A Retrieval-Augmented Generation (RAG) system that allows users to upload PDFs and query them using a local LLM. Built with FastAPI, LangChain, Ollama, and a custom JavaScript interface.

Comment

AI Document Intelligence Assistant

A full-stack Retrieval-Augmented Generation (RAG) system that allows users to upload PDF documents and ask questions about them using a local LLM.

The system performs semantic search over document embeddings and returns grounded answers with citations.

Demo

Demo

Interface

Dashboard

Features

• Upload and index PDF documents
• Automatic chunking and embeddings
• Semantic vector search
• Grounded answers with citations
• Local LLM inference using Ollama
• FastAPI backend
• Custom JavaScript frontend UI

Architecture

PDF → Loader → Chunking → Embeddings → Vector DB ↓ Semantic Search ↓ Local LLM (Ollama) ↓ Answer + Citations

Tech Stack

Backend

  • Python
  • FastAPI
  • LangChain
  • ChromaDB
  • Ollama
  • Llama 3

Frontend

  • HTML
  • CSS
  • JavaScript

Project Structure

app/ ├── api/ ├── services/ ├── models/ ├── db/

frontend/ ├── index.html ├── app.js ├── styles.css

data/ ├── uploads ├── vectordb

assets/ ├── dashboard.png ├── demo.gif

Running Locally

Clone the repo git clone https://github.com/bdcreativesystems-star/ai-document-intelligence

Share this project:

Updates