Feature
Automatic is not the same as out of your hands. Every decision shows on the transcript
Most automatic editors hand you a rendered file and one lever, regenerate. This one makes every first decision itself, then shows each one on the transcript of your take, where reversing it means pointing at it. Here is what gets decided for you, where you read it, and what a finished pass costs.
- Decisions hidden from the transcript
- 0
- Transition styles it chooses from
- 5
- The whole automatic pass, one take
- 100 CR
- Per minute of exported video
- 20 CR
What gets decided for you before you touch anything
Upload one spoken take of up to three minutes. The machine transcribes it with word-level timing, then makes the whole first pass of decisions.
Where the cuts fall, always at sentence ends. What gets tightened, so the dead air between sentences and the false starts go. Which words need footage, and which clip covers each of them.
How the captions are timed, word by word, from your own audio. What the opening headline should say. How each clip enters, from five transition styles: Cut, Whip, Punch, Glitch or Sweep.
That pass is 100 credits, and it returns a watchable video rather than a project. The question this page answers is what happens on the day you disagree with one of those decisions.
The automatic pass, and where you re-enter it
The lever this flow does not contain is regenerate. Fixing one decision never re-rolls the rest.
Upload one take
3 min ceiling, speech required
The machine decides
Cuts, clips, captions, headline
Decisions land on the transcript
Read every one
You point
Swap, restore, re-pace
Recompute and export
20 CR a finished minute
A render you cannot read is a slot machine with an export button
The category habit is to hide the decisions inside the render. You get a finished file, and when something is wrong you get one lever: generate it again.
That lever re-rolls everything. The clip you hated goes, and so does the cut you liked, because the machine cannot be told which was which.
So you sit pulling the lever, trading one stranger's edit for another, and the tool that promised to save an afternoon quietly takes one.
Automatic means doing the work for you. Opaque means refusing to show you the work. The category sells the first and too often ships the second, and the difference only appears on the day you disagree with the machine.
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
Here every decision lands on the transcript, where you can point at it
The first cut arrives displayed on the words of your take. The lines the cut dropped stay readable, struck through, so you can see what went and put a line back.
Each placed clip sits on the exact phrase it covers. The drafted headline is there to keep or rewrite. The pace, the caption pack and each transition are visible settings, not baked outcomes.
Disagreeing is pointing. Swap the clip in one spot, or search millions of free clips for a better one. Restore the line. Move the emphasis. Change the pace and the whole read re-cuts.
Nothing re-rolls. Your instructions are applied to the original take and the cut is recomputed, so fixing one decision cannot disturb the nine you agreed with.
The rules are stated, so the output is predictable before you pay
Cuts land at sentence ends and nowhere else, because a cut inside continuous speech is a join you can hear.
Footage coverage is capped by the pace you pick: 40 percent at chill, 45 at normal, 52 at fast. The machine will not build the wall-to-wall montage, because the person talking is the reason anybody believes the video.
Captions come from one of six shipped packs, and every export is loudness normalised so the file lands at the level feeds expect.
A tool that states its rules can be predicted, and a tool you can predict is one you stop babysitting. That is the practical product of the transparency, not the ideology of it.
A one-button auto editor against a visible one
The first two rows are real losses. They buy the honesty of the four below them.
| A one-button auto editor | Cutroom | |
|---|---|---|
| Works on footage with nobody speaking | A montage needs no voice | The transcript is the interface, so speech is required |
| Landscape and square masters | Many render several shapes | One shape: 9:16 vertical |
| Shows its decisions before you export | The render is the first thing you see | Every decision sits on the transcript |
| Fix one decision, keep the rest | The only lever is regenerate | Point at it and the cut recomputes |
| States its cutting rules | The rules live inside the model | Sentence ends, capped coverage, named packs |
| The ninth version of the same take | Nine re-rolls, nine strangers | Minutes of marking, 20 credits an exported minute |
What it will not decide, what it costs, and who should leave
It will not decide what to say. The machine cuts what you said and has no opinion about whether it was worth saying. A weak argument comes back as a tidy video of a weak argument.
The batch is 100 credits: transcription, the directing pass, the footage search. Export is 20 credits per minute of finished video, so a thirty second video is about 110 credits end to end.
Three finished videos are free inside your first seven days, no card, on the 600 credits granted at signup, and they export with a Cutroom mark across the middle. Lite is $19.99 a month for 1,250 credits, nothing on the picture, cancel in one click. Basic is $39.99 for 2,500, Premium is $79.99 for 5,000, and a $39.99 top-up adds 2,500.
And who should leave: anyone whose source is a podcast, a webinar or several files. This automates the edit of one spoken take into one vertical video. A long recording needs a clipping tool, and a multi-source edit needs a timeline.
Questions people ask
- If the editing is automatic, what is left for me to do?
- Judgement. You read the cut on the transcript, disagree with a decision or two, and point at them. Most sessions take minutes, and most of the machine's calls survive them.
- Can I see what it cut out of my take?
- Yes. Dropped lines stay readable in the transcript rather than vanishing, so the cut is an argument you can audit. If it dropped the wrong line, restore it and the video rebuilds.
- What happens when it picks the wrong clip?
- You swap it. Point at the spot, take another suggestion or search millions of free clips, and the change costs no credits. Only the export is charged, at 20 credits per finished minute.
- Does automatic mean every video comes out the same?
- The rules are the same, the inputs are not. Your take decides the words, the pace decides the density, the pack decides the look. Two takes through the same rules do not resemble each other.
- Is it the right tool for my footage?
- Only if the footage is one person speaking for up to three minutes and the deliverable is vertical. Podcasts and webinars need a clipping tool. Multi-camera or multi-source edits need a timeline. Bring the spoken take here and keep those there.
Let the machine make every first decision. Just never accept an editor that will not show you which decisions it made.