๐ LearnNest
๐ก Inspiration
We wanted to make studying smarter and safer.
PDFs are everywhereโresearch papers, textbooks, manualsโbut theyโre hard to digest.
๐ What if AI could summarize content, generate quizzes, and create personalized study notes?
๐ What it does
- ๐ Smart PDF Processing โ Upload and process PDFs, extracting meaningful content
- ๐ AI-Generated Quizzes โ Custom MCQs, True/False, and short answers using Google Gemini
- โ
Intelligent Assessment โ Evaluates answers, explains mistakes, and gives feedback
- ๐ Personalized Notes โ Highlights weak areas and strengths with study notes
๐ ๏ธ How we built it
Backend
- โก FastAPI for APIs
- ๐ค Google Gemini AI for quiz & notes generation
- ๐ฅ Firebase Firestore for user data & storage
- โ๏ธ Cloudinary for PDF storage
- ๐ Pinecone for semantic search
- ๐ PyPDF2 / PDFPlumber for PDF parsing
Frontend
- โ๏ธ React + Vite
- ๐ฆ TypeScript
- ๐จ Tailwind CSS
๐งฉ Challenges we ran into
- โ ๏ธ Token/context limits for large documents
- โฑ๏ธ Balancing speed vs. accuracy in quiz generation
- ๐ง Firebase initialization issues โ โNoneType has no attribute 'collection'โ
- ๐ก๏ธ Designing graceful fallbacks for when external APIs fail
๐ Accomplishments weโre proud of
- ๐ 99% uptime with retry + fallback logic
- ๐๏ธ Scalable architecture for multi-user + large PDFs
- โก Smart error handling with graceful degradation
- ๐ End-to-end security (auth, authz, validation)
- ๐ฏ Great UX: clean APIs โ seamless frontend
๐ What we learned
- ๐ค AI APIs need robust error handling + fallbacks
- ๐๏ธ Importance of database design for dev & prod modes
- ๐ Structured API design massively improves frontend dev speed
๐ฎ Whatโs next for LearnNest
- ๐ฑ Mobile App โ React Native
- ๐ฅ Collaboration โ shared quizzes & study groups
- ๐ Advanced Analytics โ ML-driven study path recommendations
- ๐ Multi-language support โ PDFs in many languages
โจ Tagline for Devpost:
๐ โLearnNest: From PDFs to Quizzes & Notes โ Smarter Studying Powered by AI.โ

Log in or sign up for Devpost to join the conversation.