PDFLyzer

Chat with your PDFs. Get answers grounded in your documents.

PDFLyzer is an AI-powered document intelligence application that lets you upload PDF documents and interact with their contents through context-aware conversations.

Powered by Retrieval-Augmented Generation (RAG), PDFLyzer retrieves relevant information from your documents to generate meaningful, context-grounded responses with source references.

Features

• Document Upload – Upload and manage PDF documents.
• AI-Powered Conversations – Ask questions and interact with your documents using natural language.
• Semantic Search – Retrieve relevant document content based on meaning rather than exact keyword matches.
• Context-Grounded Responses – Generate answers using retrieved document context.
• Source References – Trace AI-generated responses back to their relevant document sources.
• Streaming Responses – Receive generated responses in real time.
• Asynchronous Processing – Process documents in the background without blocking the application.

Project Structure

PDFLyzer is organized into two independent applications:

PDFLyzer/
├── frontend/       # React + TypeScript + Vite
│   └── README.md
│
├── backend/        # Python + FastAPI
│   └── README.md
│
└── README.md

Getting Started

To get started, navigate to the respective application directory and follow its setup instructions.

Component	Description	Documentation
Frontend	Web application built with React, TypeScript, and Vite.	Frontend README
Backend	REST API and background processing powered by FastAPI.	Backend README

Prerequisites

• Node.js and pnpm
• Python 3.14+ and uv
• MongoDB
• Redis
• Google Gemini API key

Refer to the individual application READMEs for installation, environment configuration, and development instructions.

Tech Stack

Frontend

• React
• TypeScript
• Vite
• Tailwind CSS
• TanStack Query

Backend

• Python
• FastAPI
• MongoDB and Beanie
• Redis and ARQ
• Sentence Transformers
• Google Gemini
• PyMuPDF