01
Drop in a long video
Upload an MP4 or paste a YouTube URL. ClipEngine pulls the source and validates it before anything else touches it.
Drop a YouTube URL or upload an MP4 — ClipEngine cuts the moments worth sharing in a pack of clips ready for every feed.
Sign in to make your first pack — it takes 30 seconds and your first few clips land in your Library automatically. Want to chat with an AI editor instead? Studio is the chat surface.
Podcast couch, sports broadcast, cooking demo, lecture hall, concert footage, raw vlog dump — ClipEngine finds the moments worth sharing in every kind of video.
Upload an MP4 or paste a YouTube URL. ClipEngine pulls the source and validates it before anything else touches it.
Auto-transcription + a scoring pass surface the windows most likely to land. You re-roll the list and drop the ones that don't.
Face, body, combo, horizontal. Each variant gets its own sidecar so you can re-render against the original cuts at any point.
Studio is a chatbot, not a dashboard. Tell it what you want — “tighter on the face,” “1:1 for IG,” “swap to glass captions, render 5 variants” — and the agent picks the right edit and runs it. Anything you can do in Studio you can also drive from your own scripts via the API.
No drag-rectangles, no timeline panels, no “export preset” menus. You describe the cut; the agent does the rest. Trim, crop, transcribe, burn-in captions, generate a hook, render in four aspects — one prompt, parallel jobs, finished clips in your library when it’s done.
Every render remembers the choices that made it. A follow-up like “same clip but vertical” reuses everything except the one thing you changed. Caption restyles finish in seconds, not minutes.
The chatbot is just the surface. Every command is a real action you can also drive from your own scripts.
Same prompt, same source, same output. Re-render with one new word; the previous choices stay pinned.
“Trim, vertical crop, captions, hook, render in 3 aspects” → one prompt, one job, four MP4s.
Drive an entire dailies workflow by prompt — transcode, index a four-hour interview by what was said, audit codec drift, sketch a rough cut. Then take the timeline into Resolve.