Inspiration
YouTube is the world’s biggest free classroom—but watching a 60-minute lecture to capture a few key insights wastes time. As developers, students, and content creators ourselves, we wanted a tool that turns any YouTube video into concise notes and searchable transcripts without copying links or juggling tabs. The RevenueCat Ship-a-ton 2025 hackathon gave us the perfect push to build it.
What it does
NoteTube AI lets you search YouTube directly inside the app, pick a video, and instantly generate:
AI-powered summaries – clear, bite-sized takeaways.
Smart notes – structured outlines of long lectures or tutorials.
Full transcripts with timestamps – jump straight to the exact moment a topic appears.
Keyword highlight & search – find and quote ideas instantly.
Users can save or share their notes for study, research, content creation, or collaboration.
How we built it
Frontend: Flutter with a clean, modular architecture for Android & iOS.
Backend/AI: FastAPI + OpenRouter for LLM calls (GPT-4-Turbo & Claude).
Subscriptions & Paywall: RevenueCat SDK with a custom Swift/Flutter paywall for in-app purchases.
YouTube Data: Official YouTube Data API v3 for in-app video search and metadata.
Storage & Analytics: Firebase (Auth, Firestore, Crashlytics, Remote Config, Analytics).
We followed a serverless-lean model so the heavy lifting—summarization, transcript parsing, and multilingual processing—happens in the cloud for scalability.
Challenges
Real-time transcript alignment – ensuring timestamps stay accurate across languages and video speeds.
RevenueCat integration – syncing cross-platform entitlements while supporting promo codes during the hackathon window.
Latency & cost management – balancing high-quality LLM output with speed and affordable token usage.
Accomplishments
First public release (v1.0.6) on Google Play in under 6 weeks.
End-to-end AI workflow with zero manual copy-paste steps.
A smooth paywall experience powered by RevenueCat subscriptions.
What we learned
Hackathon speed demands ruthless prioritization. We learned to cut non-MVP features (like offline summaries) and focus on a single, delightful core: “Search → Select → Summarize.”
What’s next
iOS launch and macOS desktop client.
Collaborative “shared notes” spaces for study groups.
Fine-tuned domain models for education, business, and medical content.
Log in or sign up for Devpost to join the conversation.