How CutGPT decides what to cut

CutGPT reads your take word by word and decides what to keep. The web app and the iPhone app use the same rules.

What happens to your video

1. Transcribe. It writes down each word with its timing, matched to your voice.
2. Split. It breaks your speech into phrases at pauses and sentence ends.
3. Cut. It removes the mistakes listed below.
4. Tighten. It joins what's left and trims dead air.
5. Caption. It adds captions.

What gets cut

• Retakes. If you say a line again within the next few phrases, the earlier try counts as an abandoned take. CutGPT keeps the last take.
• Restarts. When a phrase is just a cue, like “wait”, “sorry”, “scratch that” or “let me try that again”, CutGPT removes it and the attempt just before it.
• Fillers. Um, uh, erm, hmm and the like. See the full list.
• Slates. An “Okay” before your first real line, and a “cut”, “done” or “stop” after your last.
• Dead air. Longer silences are trimmed down to a breath. Your cut pace sets how long a pause can stay.

Check what was cut

In the web app, the result page lists every cut with its reason: retake, restart, slate, filler, dead air or cut by you. Click a cut to jump to it. If CutGPT got one wrong, fix it in the editor.

The page also estimates the time you saved: 1.5 times your raw footage length for scrubbing, plus 10 seconds per cut and 1 second per caption word. It's a deliberately conservative estimate, not a measurement. The money figure is based on Your time is worth in Settings.