Workflow
From long recording to vertical short, entirely in your browser
The repeatable six-step workflow we use on our own footage: cut, tighten, caption locally, check safe zones, export. Every claim below is demonstrated by the unedited capture and the real output file.
1 · Get one vertical file into the editor
Drag a clip straight from your recordings folder into the Assets panel — the browser reads it locally, nothing uploads. The timeline auto-expands to fit your media, and the first clip lands on V1 at the playhead.
2 · Cut to the hook
Scrub with the timeline ruler, park the playhead where the payoff starts, and press S to split everything at the playhead. Delete the dead head with the Delete key. Trims are source-aware: dragging a clip's left edge pulls earlier source material when it exists and never past it.
3 · Tighten with Auto-Adjust
The Auto-Adjust button ripples every track left to close the gaps your deletions left behind, keeping order intact. Caption word timings shift with their clips so audio and captions stay in sync.
4 · Caption locally
Open AI Tools → Captions → run transcription with Timeline Media as the source. Whisper runs on your device (WebGPU when available, WASM otherwise) and the word-level captions land on T1 automatically. Pick a kinetic style, then fine-tune individual words — rename, retime, delete — in the properties panel.
5 · Check platform safe zones
Toggle Safe Zones with TikTok, Reels or Shorts selected. The overlay simulates each platform's UI chrome and warns when your captions collide with it; Auto-Fix moves captions into the clear band.
6 · Export and archive
Export renders MP4 or WebM at your chosen resolution and frame rate in real time, watermark-free. Save the .json project backup next to your footage — it restores the whole edit, including timing.
First-hand captures
Recorded from the production build of this site: a real browser session driving the real editor, with no cuts or retakes. Output files below were produced by the editor's own export engine.
Supported formats
- Works with any MP4, MOV or WebM recording your browser can decode; vertical, square and widescreen frames
- Captions: word-level timestamps from on-device Whisper (tiny.en ~39 MB or base ~73 MB model, downloaded once per browser)
- Export: MP4 (H.264) where supported, otherwise WebM; 720p–4K, 30/60 fps
- Timeline state persists in browser storage; .json project backups are portable
Known limitations
- Whisper transcription quality depends on audio clarity; heavy music beds over speech reduce word accuracy
- The first transcription downloads the model — budget for that one-time download on slow connections
- Exports run in real time and pause if the tab loses visibility
- Multi-clip silence-aware jump cutting is engine-assisted, not a guarantee — review cuts before export
Open the editor and run these six steps on one of your own recordings — no account needed. For the text and layering half of the workflow, see the text and layers tutorial.