Bookified

About the Project

Bookified — a full-stack AI platform that lets users have real-time voice conversations with their books. The app transforms uploaded PDFs into interactive entities, the user speaks, the AI listens with full awareness of the book's content, and responds naturally using a chosen voice persona, complete with live transcripts and on-demand summaries.

Role: Fullstack Developer

What I built:

• Next.js 16 and TypeScript full-stack app with MongoDB and Mongoose for storing user libraries, book metadata, and conversation transcripts

• PDF ingestion pipeline with automated text extraction, intelligent chunking, and high-dimensional embeddings for precise context retrieval during voice sessions

• Real-time voice conversation engine using Vapi — handling turn-taking, low-latency audio streaming, and back-and-forth dialogue grounded in uploaded book content

• AI voice persona system powered by ElevenLabs — users preview and select from distinct, lifelike voices before starting any session

• Smart summary and insights feature for extracting chapter-level takeaways and deep-dive explanations on demand

• Auto-generated session transcripts so no insight from a conversation is ever lost

• Full library management interface with high-performance search across personal and global uploads

• Clean, modular UI built with shadcn/ui and Tailwind CSS, with secure auth and subscription management via Clerk