VantraClip
VantraClip is an AI-assisted video clipping platform in active development. It processes long videos into shorter clips with highlight detection, captions, face-aware reframing, branding, and multiple export formats.
Challenge
Coordinate a long-running media workflow that can accept a source video, identify useful moments, create readable captions, reframe footage, and export clips without blocking the main application.
Solution
The product combines a React and TypeScript client with Express services, PostgreSQL, S3-compatible storage, and FFmpeg processing. API and worker responsibilities remain separable while sharing the same data and storage layer.
Key decisions
Keep API and processing-worker roles separable for long-running video jobs.
Snap clip boundaries to complete spoken thoughts instead of forcing fixed cuts.
Use signed browser uploads so storage credentials remain server-side.
Outcomes
Built the core workflow for uploads, transcription, highlight selection, captions, reframing, and multi-format exports.
Established the full-stack foundation for authentication, projects, storage, processing jobs, and delivery.
Next steps
Continue refining processing reliability, output quality, and the end-to-end product experience.