SnapLearn — Multimodal AI Study Buddy

InspirationMillions of Indian students study from textbooks but get stuck on diagrams they can't understand. Private tutors are expensive. I wanted to give every student a free AI study buddy that sees what they see.What it doesSnapLearn is a multimodal AI study companion. Snap a photo of any textbook diagram or page:- Image AI analyzes it — topic, subject, diagram labels- Explains simply — in English or Hindi, with voice readout- 3 difficulty levels — ELI5, class level, advanced- Quiz agent — interactive MCQs with explanations- Flashcard agent — tap-to-flip revision cards- Voice Q&A — ask questions by speaking- Progress dashboard — streaks, scores, topics mastered- Export notes — download as markdownThree real modalities: image in, voice in/out, text in/out.How I built it- Frontend: React + TypeScript + Tailwind CSS- Backend: Express + TypeScript, OpenAI-compatible vision API- Voice: Web Speech API (free, no key needed)- AI: Pluggable vision backend (Nebius/Qwen-VL); honest demo mode for judging- Testing: 11/11 E2E tests passingChallengesMaking the demo honest: instead of hallucinating analysis for any image, demo mode clearly labels sample content and guides users to connect a real vision key. The difficulty selector actually changes explanation depth.What's next- Real-time vision API as default- More subjects, NCERT chapter mapping- Spaced repetition for flashcardsAI DisclosureBuilt with AI assistance (Muse). All code reviewed and tested by the developer.

Built With

• express.js
• node.js
• react
• tailwind-css
• typescript
• web-speech-api