Several finished versions from one recording
Post the same talk to TikTok and LinkedIn and each wants something different. After one analysis, Cutroom AI lays out several finished versions side by side, so you pick instead of re-editing.
Updated November 12, 2026
The versions you get
Open "See versions" from a finished talking-head edit. Depending on the recording, you get three to six:
- Natural cut with plain captions: relaxed pacing, the safe choice for a calm explainer.
- Tight cut with keyword captions: fast pacing, punch-in zooms on jump cuts, keywords in your accent color.
- Hook first with karaoke captions: the opening sentence of your strongest stretch moved to the front, so the video starts on its best line.
- A 30 s clip and a 15 s clip with boxed captions and the clip's hook line on screen. Each appears only when the edit is long enough to cut one from.
- Long form, no captions: a full natural cut in the shape you filmed, offered when you filmed horizontally or the talk runs past a short format.
Pick the destination first
"Made for" sets the shape: 9:16 for TikTok, Reels and Shorts, 16:9 for YouTube and X, 4:5 for LinkedIn, 1:1 for Square. Short clips are kept to the length range each feed suits.
Compare, pick, export
Versions play side by side, so you can hear the natural and tight cuts back to back and decide which pace suits the talk. Press Pick on the ones you want, then "Export picked" to render them together, or "Save to projects" to keep working on them.
Each version is its own project made from the same takes. Editing one never changes the others, so you can trim the hook-first version harder without touching the natural one.
An example
Say you filmed a 90-second explainer on how to price a first freelance job. The natural cut is the one for your website, where people are already listening. The tight cut, with keyword captions on the numbers, suits a Reel. Hook first opens on "Charge for the outcome, not the hour", a line from your second paragraph, which gives a scroller a reason to stay. The 30-second clip covers just the pricing rule, with that line on screen. You pick three, export them together and post each where it fits.
How to use it
- 01Import and let the cleanup finish. Versions need a talking-head edit with a transcript.
- 02Press "See versions".
- 03Choose a destination under "Made for".
- 04Play versions side by side. If you start more than can play at once, the first one stops.
- 05Pick the ones you want and press "Export picked".
- 06Open any version as a project to adjust it further.
What it won't do
It doesn't write a new hook. The hook-first version reorders your own sentences, because AI-written hooks and titles are switched off.
It doesn't post to your accounts or test which version gets more views. You decide what goes where.
Montage projects and recordings with no speech don't get versions. Versions are built from the transcript.
When another approach is better
If a video only ever goes to one place, skip versions and export the main edit. If you want many short clips from a long talk, rather than a few cuts of one short video, use the clip finder in the Clips tab. It ranks every candidate stretch across the whole recording.
Questions
Is each version a separate project?
Yes. Each one is saved as its own project from the same takes, so editing one leaves the others alone.
Why don't I see a 30-second clip?
Short clips appear only when the edit is long enough to cut one from. A 20-second video won't get a 30-second clip.
Can I export several versions at once?
Yes. Pick the ones you want and press "Export picked". They render one after another on your device.
Keep reading
- FeatureFind the clips worth posting in a long recording
- FeatureAuto captions timed to every word you say
- FeatureEdit a full-length talk without chopping it into shorts
- AnswerWhat aspect ratios can I export?
- GuideHow to make a vertical video from horizontal footage
- Who it's forFor founders who are also the video team
Try it on your own footage.
The editor opens in your browser with no account, and the sample project doesn't use an edit.