📖 ChatBook – Real-Time AI-Powered Book Chat

✨ Inspiration

We’ve always believed in the power of books to spark conversations, share knowledge, and change perspectives. But while platforms exist to discuss books or access summaries, none offer a unified space where readers can collaboratively chat, interact with the book directly using AI, and receive intelligent assistance — all in real time.

ChatBook was born from that gap.

We wanted to build a platform where users could:

Upload any book and talk to it directly using AI
Engage in real-time group chat about the content
Get instant AI-powered answers if something didn’t make sense
Explore knowledge collaboratively, beyond traditional static reading

With Auth0 for secure identity, MongoDB Atlas for smart storage, and Gemini + GPT-4 for intelligent understanding, we set out to redefine the way people read, learn, and discuss books together.

🛠️ What We Built

ChatBook is a full-stack AI-powered web app that enables users to:

🧠 Upload PDF or TXT books
🔍 Perform semantic search within the content
💬 Engage in real-time group chats
📚 Chat with specific books using vector-based RAG pipelines
🤖 Ask general questions to Gemini AI
🔐 Log in securely using Auth0
📂 Upload and interact with multiple books per session

🧪 How It Works

🧬 AI Integration

Gemini AI (Google) powers the general-purpose assistant.
GPT-4 is used for vector embeddings and semantic understanding.
A custom Retrieval-Augmented Generation (RAG) pipeline was built using LangChain, allowing intelligent, context-aware answers from uploaded books.

📚 Book Interaction

When a user uploads a book, it is parsed using PyPDF and chunked into vector embeddings.
These are stored in MongoDB Atlas and indexed for semantic retrieval.
Users can select a book from a sidebar and ask context-specific questions to the AI.

💬 Real-Time Chat

Built with Flask-SocketIO and Gevent, the group chat allows users to communicate instantly.
WebSocket-based messaging is supported using Socket.IO for two-way real-time communication.
AI assistance is context-aware and available in both book-specific and general chat modes.

🔐 Authentication

We used Auth0 for secure login, route protection, and token-based session management.
User identity is tied to uploaded content, enhancing the personalized experience.

⚙️ Tech Stack

Backend

Flask (Web framework)
Flask-SocketIO (WebSocket support)
Gevent (Async I/O)
PyMongo (MongoDB access)
LangChain (RAG orchestration)
OpenAI API (GPT-4 Embedding)
Gemini API (LLM Assistant)
Auth0 (Authentication)
python-dotenv (Config management)

Frontend

HTML, CSS, JavaScript
WebSocket + Socket.IO

Database

MongoDB Atlas
Vector search for semantic content understanding

DevOps

Auth0 integration
Environment-based config
SSL (for secure local communication)
Git & venv for development

🧠 Challenges We Faced

Real-Time Communication

Challenge: Syncing AI responses and user chat across multiple users
Solution: Used event-namespacing with Socket.IO and scoped room management

Book Processing & AI Context

Challenge: Handling large files and ensuring meaningful context for AI
Solution: Implemented PDF chunking, semantic vectorization, and a LangChain-based RAG flow

Performance Bottlenecks

Challenge: Maintaining response time with growing vector DB
Solution: Optimized caching and minimized redundant calls

Secure Access

Challenge: Simplifying yet securing the auth workflow
Solution: Used Auth0’s hosted UI and rules to manage sessions and scopes

💡 What We Learned

This project was a deep dive into:

Real-time web systems and WebSocket architecture
Language model orchestration with LangChain
Secure file and user session handling with Auth0
Building RAG pipelines and working with vector embeddings
Designing for collaborative, intelligent user experiences

We also explored how powerful AI can become when paired with social features — like conversation, collaboration, and knowledge sharing.

🔮 What’s Next

✅ Support for EPUB and DOCX formats
✅ AI-powered book recommendations
✅ Multi-language support
✅ Exportable chat summaries
✅ Collaborative reading rooms with moderation
✅ Hosting for public access + demo site

🌍 Social Good Impact

ChatBook fits the Hacks for Social Good track by:

Making knowledge more accessible and conversational
Helping students and lifelong learners explore books collaboratively
Offering AI assistance to those with learning differences or limited access to tutors
Creating a platform that encourages curiosity, reading, and critical thinking