Passive watching, meet active mastering.
NovaLecture AI turns any video or YouTube lecture into a complete suite of study materials—summaries, flashcards, PPT presentations, and animated videos.
Built with Amazon Nova 2 Lite for the #AmazonNova Hackathon.
NovaLecture AI is your lecture, simplified. Get instant AI summaries, flashcards, PPTX presentations, and animated videos powered by Amazon Nova. Master any subject in minutes.
- ⚡ Instant Transcription — High-speed audio extraction and transcription using
faster-whisper. - 📝 AI Summarization — Deep reasoning summaries and key concept extraction powered by Amazon Nova 2 Lite.
- 🗂️ Active Recall Kit — Automatically generated MCQs, digital flashcards, and short answers.
- 📊 Visual Study Kit — Programmatically generated PowerPoint presentations (.pptx).
- 🎞️ Animated Explainers — Narrated, scene-by-scene 60-second summary videos (.mp4).
- 🙋 Interactive Tutor — Ask questions about the lecture and get answers based strictly on the transcript.
┌─────────────────┐ ┌──────────────────┐ ┌─────────────────┐
│ │ │ │ │ │
│ React App │◄─────►│ Main API │◄─────►│ Amazon Bedrock │
│ (Port 5173) │ │ (Port 8000) │ │ Nova 2 Lite │
│ │ └────────┬─────────┘ └─────────────────┘
└─────────────────┘ │
│ /generate-video
▼
┌──────────────────┐
│ Video Generator │
│ (Port 8001) │
└──────────────────┘
- AI Model: Amazon Nova 2 Lite (via Amazon Bedrock Converse API)
- STT:
faster-whisper(tiny.en for demo speed) - Backend: Python 3.14, FastAPI, boto3
- Media:
python-pptx,MoviePy,gTTS,Pillow - Frontend: React (Vite), Lucide Icons, Glassmorphism Design
- Tools: yt-dlp (YouTube support), FFmpeg (Media processing)
cd backend
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
cp .env.example .env # Add your AWS Credentials
python main.pycd video_generator
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
python main.pycd frontend
npm install
npm run devThe genesis of NovaLecture AI came from a simple observation: students are drowning in content. We built this tool to bridge the gap between passive viewing and active recall.
We modeled information density as: $$ I = \sum_{i=1}^{n} \frac{T_i}{V} $$
By leveraging Nova 2 Lite's low-latency reasoning, we can process a 60-minute lecture into an interactive kit in under 20 seconds.
Built with ❤️ for the #AmazonNova Hackathon