Your vision. In focus.
Good stories start here.
Import a photo or video to begin.
Your originals stay untouched.
Director · plan your first cut
Local assembly uses favorites, quality scores and detected scene boundaries when available. It cannot recognize people or understand your story.
Optional hosted AI uses Tero8’s shared free allowance. Review its proposal before applying.
Your sequence
Trim each source, then add it. Photos, video and titles can share a sequence. Preview before export; recording runs in real time, up to 5 minutes.
Favorites and analyzed segments guide local assembly. It fills up to your target length from the available material.
Music, voiceover & sound
Creates up to 60 seconds of original synthesized music from your target length. Beat snapping uses the entered BPM and timeline zero as its grid; it does not detect beats in an imported song. Recheck captions after snapping.
Signal-based ducking reacts to clip volume, not speech recognition. Preview the mix. Import audio you have permission to use.
Captions & chapters
English speech recognition downloads a model on first use and runs locally. It processes up to 90 seconds of selected audio, or the selected video trim range. Your browser must be able to decode its audio. Check every word and timestamp; silence can produce incorrect text.
For video transcripts, each kept cue becomes a source-range cut. Editing caption text does not change the source audio. Creating this sequence replaces the current cut; Undo restores it.
Video output
Exports are SDR. Keep this tab visible. High resolution and frame rates need more memory and may drop frames on slower devices.