Text-based editing
Once your audio is transcribed, the transcript becomes an editing surface: change the text and the timeline follows.
Remove words
Select words or sentences in the transcript and remove them — the corresponding video and audio are cut out, and the timeline closes the gap. It's the fastest way to:
- Cut filler ("um", "you know", false starts) from an interview.
- Drop a whole tangent without scrubbing for its in and out points.
- Tighten a talking-head take down to the script that survived.
The transcript stays word-accurate, so cuts land on word boundaries, not somewhere in the middle of a syllable.
Remove silence
Remove silence finds every silent gap in the selected clips and ripple- deletes them in one pass — fourteen gaps totalling 22 seconds becomes one click. Thresholds keep natural breathing room; it trims dead air, not pacing.
Cut to the beat
Beat detection marks the beats of a music track. Snap your cuts to the markers, or ask the agent to "cut the b-roll on the beat" and it uses the same markers.
Where this shines
Text-based editing and the timeline aren't separate modes — a transcript cut is a timeline cut. Rough the story in text, then fine-tune the frames on the timeline.