Drop a podcast or video link. An agent crew finds the best moments, cuts vertical clips, burns karaoke captions, and ships them to your feed.
Paste a link and a pipeline of agents does the rest:
- Ingest the source video (yt-dlp)
- Transcribe with word-level timestamps (ElevenLabs Scribe)
- Score moments and pick the best clips (gpt-5.6-sol)
- Render 9:16 clips with animated captions (Remotion)
- Publish the top clip to a real social account
- Unlock the full batch behind a checkout (Dodo Payments)
| Layer | Tech |
|---|---|
| Frontend | Next.js 15, React 19, Tailwind, Framer Motion |
| Video | Remotion 4 |
| Model | gpt-5.6-sol via the Vercel AI SDK |
| State and paywall | Convex |
| Pipeline | RabbitMQ, Redis, MinIO, Postgres workers |
| Transcription | ElevenLabs Scribe |
| Payments | Dodo Payments |
See RUN.md for the full setup and run guide.
npm install
npm run dev # app on http://localhost:3000
node --import tsx worker/clip/dispatcher.ts # the pipeline workerBackend services run via docker compose up -d (RabbitMQ, Redis, MinIO, Postgres).