📚 Bookies

An AI-powered Book Mentor that transforms books into interactive conversations using Retrieval-Augmented Generation (RAG), semantic search, and Google Gemini.

Bookies enables readers to upload books and interact with them through natural language. Instead of relying on the language model's general knowledge, Bookies retrieves the most relevant passages from the uploaded document and uses them to generate grounded, context-aware responses.

---

Live Demo

Application: https://bookies-nu.vercel.app

Note: The backend is hosted on Render's free tier and may take up to a minute to wake up after inactivity.

Features

• JWT Authentication
• Multi-format document upload (PDF, EPUB, DOCX, TXT)
• Retrieval-Augmented Generation (RAG)
• Semantic search with Sentence Transformers
• Google Gemini 2.5 Flash integration
• Book-specific workspaces
• Persistent chat sessions
• Markdown responses
• Production deployment on Vercel, Render and Neon PostgreSQL

Technology Stack

Layer Technologies

---

Frontend React, TypeScript, Vite, Tailwind CSS Backend Django REST Framework Database PostgreSQL (Neon) AI Google Gemini 2.5 Flash Embeddings all-MiniLM-L6-v2 Deployment Vercel, Render

Architecture

Project Structure

Bookies/
├── backend/
├── frontend/
├── assets/
├── README.md
├── ARCHITECTURE.md
├── API.md
└── DEPLOYMENT.md

Running Locally

Backend

cd backend
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt
python manage.py migrate
python manage.py runserver

Frontend

cd frontend
npm install
npm run dev

Environment Variables

Backend:

SECRET_KEY=
DEBUG=False
DATABASE_URL=
GEMINI_API_KEY=
ALLOWED_HOSTS=
CORS_ALLOWED_ORIGINS=

Frontend:

VITE_API_BASE_URL=http://localhost:8000

Engineering Highlights

• End-to-end RAG pipeline
• Semantic retrieval
• Context-grounded LLM responses
• Modular Django architecture
• Ownership-based authorization
• Production deployment

Roadmap

Completed

• Authentication
• Document ingestion
• RAG
• Chat sessions
• Deployment

Planned

• Streaming responses
• OCR support
• Hybrid search
• Book summarization

Documentation

• ARCHITECTURE.md
• API.md
• DEPLOYMENT.md

Author

Prayag Raj Shrivastava

• GitHub: https://github.com/Prayag1-tecj
• LinkedIn: https://www.linkedin.com/in/prayag-raj-shrivastava-129875293
• Email: prayag.raj.9887@gmail.com

License

MIT License.