auth screen
pdf upload
pdf preview
results
quizzes
analogy
history

Academic PDF Companion

Inspiration

I've always found academic PDFs hard to engage with. Whether it was dense research papers or scattered lecture notes, the learning process often felt inefficient. I kept wondering: what if an AI could help break it down the way a really good tutor would? That curiosity sparked the idea for Academic PDF Companion — a tool that could make academic reading more digestible, interactive, and personal.

What it does

Academic PDF Companion lets users upload any academic PDF — such as a research paper, textbook chapter, or lecture notes — and transforms it into a structured learning experience. It provides:

Clear, simplified explanations
Quiz questions with answers hidden until revealed
Glossaries of key terms
Analogy-based explanations for deeper conceptual understanding
Page-by-page breakdowns for focused learning
Real-time PDF preview
Chat with document using session-based memory
Option to export quizzes as printable sheets
User login and upload history via persistent accounts
Light/dark theme toggle for comfortable reading

How we built it

Frontend: Built with React and styled using Tailwind CSS for a modern, responsive interface
Backend: Node.js and Express handle PDF uploads, user accounts, and secure AI requests
AI Integration: Google Gemini API used for all content generation (summaries, quizzes, explanations)
PDF Handling: Base64-encoded uploads, preview rendered using react-pdf, export with jspdf
Auth & Sessions: Google OAuth for login, session-based memory for chat, persistent user data in Firestore

Challenges we ran into

Designing prompts that consistently return usable output from Gemini
Structuring output clearly across document types (research papers vs. lecture notes)
Preventing hallucinations while encouraging useful analogies
Building a smooth, responsive UI that works well across formats and devices
Handling large and scanned PDFs with limitations in browser memory

Accomplishments that we're proud of

Built a complete full-stack app that actually helps users learn better
Integrated chat functionality with persistent document memory
Enabled quiz export and printable assessment sheets
Implemented user authentication and history tracking
Delivered a clean, accessible UI with light/dark themes and real-time document preview

What we learned

Advanced prompt engineering with Google Gemini
Securing AI interactions and API keys using a Node.js backend
Efficient PDF processing in the browser (Base64 handling, previews, exports)
Designing a feature-rich app under tight deadlines with strong UX focus
Implementing user session storage, authentication, and protected API routes

What's next for Academic PDF Companion

With the foundation in place, I'm excited to focus on:

Supporting scanned PDFs and handwritten notes using OCR
Adding automatic topic tagging and concept maps
Integrating voice input and AI-generated audio summaries
Enhancing multilingual support with auto-detection and regional fluency
Building a teacher dashboard for creating and assigning AI-generated assessments

The vision is to make this not just a tool for students, but a learning companion for anyone working with academic content.