Transform any YouTube lecture (or uploaded audio/video) into a complete AI-powered study companion — summaries, quizzes, flashcards, mind maps, and more.
| Analyze a lecture | Interactive quiz |
|---|---|
 |
 |
| Mind map & knowledge graph | Learning dashboard |
|---|---|
 |
 |
| Flashcards & spaced repetition | Landing page |
|---|---|
 |
 |
| Feature | Description |
|---|---|
| 📖 AI Summary & Overview | Comprehensive summary with key takeaways extracted by Gemini AI |
| 🗺️ Chapter Timeline | Clickable AI-generated timestamps — seek to any section instantly |
| 🔥 Key Moments | AI-identified highlights: insights, definitions, examples, turning points |
| 🧠 Interactive Quiz | 8–12 multiple-choice questions with explanations + video timestamps |
| 📝 Diverse Exercises | Fill-in-the-blank, True/False, Matching, and Short Answer exercises |
| 📇 Flashcard System | In-app spaced-repetition study mode + export to Anki (TSV) or CSV |
| 📜 Transcript Search | Full transcript with live sync highlighting, search, and click-to-seek |
| 💬 AI Chat | Ask follow-up questions about the lecture with full context |
| 🌍 Multilingual | Output in 7 languages: Vietnamese, English, French, German, Japanese, Korean, Chinese |
| Feature | Description |
|---|---|
| 📅 AI Study Plan | Personalized learning schedule generated by AI based on quiz scores, SRS data, and video history |
| 🧠 Mind Map | Visual mind map generated from lecture content |
| 🕸️ Knowledge Graph | Interactive D3.js graph connecting concepts across videos |
| 📋 Multi-Video Exam | Generate exams from multiple analyzed videos with history tracking |
| 📚 Playlist / Course Mode | Analyze entire YouTube playlists as a course with progress tracking |
| 📊 Dashboard | Learning statistics: videos analyzed, quiz scores, time spent |
| 🔥 Streak Tracking | Daily learning streak with gamification and weekly heatmap |
| 🏆 Achievement Badges | Unlock badges for milestones (first video, streak records, quiz mastery) |
| 📈 Progress Tracking | Per-video watch progress, quiz history charts, bookmark timeline |
| 🔖 Bookmarks | Mark important moments during playback with custom labels |
| 📝 Notes | Rich notes editor with auto-save, timestamp insertion, and Markdown export |
| 📊 Video Compare | Compare quiz scores and progress across analyzed videos |
| 🏷️ Video Tags | Organize videos with customizable category tags |
| 📄 PDF Export | Print-quality study guide combining summary, chapters, quiz, and notes |
| Feature | Description |
|---|---|
| Paste any YouTube video or playlist URL | |
| 📤 File Upload | Upload audio/video files directly (MP3, MP4, WAV, M4A, WebM, OGG, FLAC, AAC, MOV — up to 200MB) |
| Feature | Description |
|---|---|
| 👤 User Accounts | Email registration + login with JWT authentication |
| 🔑 Google OAuth | One-click sign-in with Google account linking |
| 🔒 Forgot Password | Secure token-based password reset via email (SMTP) |
| ☁️ Cloud Sync | Sync progress, notes, bookmarks, and history across devices in real-time |
| 📤 Share Notes | Generate shareable read-only links for notes & bookmarks |
| 👥 Study Rooms | Collaborative study rooms with invite codes, shared videos, discussion threads, and progress comparison |
| Feature | Description |
|---|---|
| 📱 PWA | Install as app on mobile/desktop with offline caching |
| 🌙 Dark / Light Theme | Toggle between premium dark and clean light modes |
| 📱 Responsive | Optimized for mobile, tablet, and desktop |
| 📜 History | Search and filter analyzed videos by name or category tags |
# 1. Clone the repo
git clone https://github.com/tridpt/LectureDigest.git
cd LectureDigest
# 2. Create .env with your Gemini API key
echo "GEMINI_API_KEY=your_key_here" > backend/.env
# 3. Build & run
docker compose up --build
# 4. Open http://localhost:8000# Requirements: Python 3.9+
cd LectureDigest
# Copy env and add your API key
copy backend\.env.example backend\.env
# Edit backend\.env → set GEMINI_API_KEY=your_key_here
# Install deps & start
pip install -r backend/requirements.txt
cd backend
uvicorn main:app --reload --port 8000
# Open http://localhost:8000# Double-click run.bat — it installs deps and starts the server automaticallyGet your Gemini API Key → aistudio.google.com/app/apikey
All config goes in backend/.env:
# Required
GEMINI_API_KEY=your_gemini_api_key_here
# Optional — Google OAuth
GOOGLE_CLIENT_ID=your_google_client_id
# Optional — Password Reset Email (SMTP)
SMTP_HOST=smtp.gmail.com
SMTP_PORT=587
SMTP_USER=your_email@gmail.com
SMTP_PASS=your_app_password
SMTP_FROM=your_email@gmail.com
# Optional — YouTube anti-block (for cloud deploy)
YOUTUBE_COOKIES_B64=base64_encoded_cookies
YOUTUBE_PROXY_URL=http://user:pass@host:portWithout SMTP configured, password reset links are printed to the server console (dev mode).
- Push your fork to GitHub
- Go to render.com/deploy → connect your repo
- Render auto-detects
render.yaml→ click Deploy - In the Environment tab, add:
GEMINI_API_KEY = your_key_here - Done! Your app is live at
https://lecturedigest.onrender.com🎉
⚠️ Free tier spins down after 15 min of inactivity — first request takes ~30s to wake up.
npm install -g @railway/cli
railway login
railway link
railway upThen set GEMINI_API_KEY in the Railway dashboard.
curl -L https://fly.io/install.sh | sh
fly auth login
fly launch
fly secrets set GEMINI_API_KEY=your_key_here
fly deploy# Build image
docker build -t lecturedigest .
# Run with env var
docker run -p 8000:8000 -e GEMINI_API_KEY=your_key lecturedigest
# Using docker compose
docker compose up --build # start
docker compose down # stop
docker compose logs -f app # view logsLectureDigest/
├── backend/
│ ├── main.py # FastAPI app — all API routes
│ ├── database.py # SQLite database (users, sync, shared notes)
│ ├── requirements.txt # Python dependencies
│ └── .env # API keys & config (never commit!)
├── frontend/
│ ├── index.html # Main SPA shell (1600+ lines)
│ ├── css/ # 35 modular CSS files
│ │ ├── base.css # CSS variables & design tokens
│ │ ├── theme.css # Dark/Light mode theming
│ │ ├── hero.css # Landing page hero section
│ │ ├── quiz.css # Quiz card styling
│ │ ├── ... # And 30+ more modules
│ │ └── mobile.css # Responsive breakpoints
│ ├── js/ # 28 modular JS files
│ │ ├── core.js # App initialization & routing
│ │ ├── analyze.js # Video analysis flow
│ │ ├── quiz.js # Quiz engine
│ │ ├── flashcard.js # Flashcard study system
│ │ ├── exercises.js # Diverse exercise types
│ │ ├── transcript.js # Transcript search & sync
│ │ ├── notes.js # Notes editor
│ │ ├── chat.js # AI chat
│ │ ├── mindmap.js # Mind map visualization
│ │ ├── knowledge-graph.js # D3.js knowledge graph
│ │ ├── exam.js # Multi-video exam
│ │ ├── gamification.js # Streaks & XP
│ │ ├── auth.js # Authentication UI
│ │ ├── db-sync.js # Cloud sync engine
│ │ ├── upload.js # File upload handler
│ │ ├── share-notes.js # Shareable note links
│ │ ├── pwa-install.js # PWA install prompt
│ │ └── ... # And 10+ more modules
│ ├── shared-notes.html # Standalone shared notes viewer
│ ├── reset-password.html # Password reset page
│ ├── manifest.json # PWA manifest
│ └── sw.js # Service worker (offline cache)
├── cloudflare-worker/ # Optional CDN worker
├── Dockerfile # Production container
├── docker-compose.yml # Local Docker dev
├── render.yaml # Render.com deploy blueprint
└── run.bat # Windows one-click start
| Layer | Technology |
|---|---|
| Backend | Python 3.13 + FastAPI + SQLite |
| AI Engine | Google Gemini 2.5 Flash (with auto-fallback to Lite) |
| Auth | JWT + bcrypt + Google OAuth 2.0 |
| Transcripts | youtube-transcript-api v1.x + InnerTube fallback |
| File Upload | Gemini Files API (audio/video transcription) |
| Video Player | YouTube IFrame API |
| Visualization | D3.js (Knowledge Graph, Mind Map) |
| Frontend | Vanilla HTML / CSS / JS (no framework) |
| PWA | Service Worker + Web App Manifest |
| Container | Docker + Docker Compose |
| Deployment | Render / Railway / Fly.io |
cd backend
pip install -r requirements.txt
pip install ruff # linter (not a runtime dep)
pytest # full suite (230+ tests)
ruff check . # lintTests run against an isolated temporary SQLite database (see
backend/tests/conftest.py) — they never touch your real data and make no
network or Gemini calls.
npm install # installs vitest + jsdom (dev only)
npm test # run Vitest once
npm run test:watch # watch modeFrontend tests focus on the XSS-escaping helpers and the Markdown study-guide
builder. They extract individual functions from the real source files (see
frontend/tests/extract.js) so a change that makes an escaper unsafe fails the
suite. .github/workflows/frontend-tests.yml runs them on every push/PR
touching frontend/.
Frontend has its own unit tests (Vitest) covering the XSS escapers and the Markdown-export builder:
npm install
npm test # Vitest (frontend)| Var | Purpose |
|---|---|
BCRYPT_ROUNDS |
Password-hash cost factor. Defaults to 12 (production). Tests set it to 4 for speed; you can lower it locally too. Never set it low in production. |
JWT_SECRET |
Signing secret for auth tokens. Required in production. |
TRUST_PROXY |
Set to true only behind a trusted reverse proxy so rate limiting reads X-Forwarded-For. |
.github/workflows/backend-tests.yml runs ruff + the full pytest suite on
every push to main and on pull requests touching backend/.
See ARCHITECTURE.md for a technical overview — request flow, backend/frontend layering, the data model, API surface, caching, and sync.
See SECURITY.md for hardening decisions (XSS escaping rules, CSP, input validation, rate limiting) and how to report a vulnerability. Read it before changing auth, input handling, or any code that renders user/AI content.
MIT