Comparison
Synthesia's viewer already pressed play. Yours is trying to leave.
Synthesia makes video for a viewer who was told to watch. Onboarding, compliance, a course module. That viewer is not the one in a feed. Here is why the two builds pull apart, what an ad costs in credits, and how Cutroom finishes one.
- what a cold viewer gives you
- 3 sec
- coverage ceiling at fast pace
- 52%
- to export another cut of the same take
- 20 CR
Synthesia is built for permanence. An ad is built to be turned off in a week.
A compliance module says the same thing to every new starter for three years, in every language the company hires in.
It has to be updatable when clause four changes. No rebooked studio, no reshoot of a person who has since left.
It has to look identical across two hundred videos. That is a template problem and a governance problem. Synthesia solves both.
An ad is the opposite object. It is disposable by design. It runs for nine days, it stops working, and the next one has to already exist.
That is what Cutroom is built around. One take on a phone becomes a finished 9:16 ad, captions burned in and footage on the words that needed showing.
The batch that does it is 100 credits and covers the transcript, the director pass and the b-roll search. Export is 20 credits per finished minute on top.
Then the second version costs 20 credits a minute, because the take is already transcribed. Volume stops being a budget conversation.
A composed presenter is exactly what a scroller has learned to skip
The polish that makes training video trustworthy is the polish that makes ad video invisible. Same finish, opposite outcome.
A viewer in a feed sorts content from advertising in under three seconds. Centred framing and even lighting are the strongest signal of advertising available.
Ad video wins on interruption, not composure. A specific number in the first line. A hesitation that reads as a person rather than a script.
Training video is judged on completion. An ad is judged on cost per result.
Nobody makes thirty compliance modules a month. Thirty ad creatives a month is a normal ask.
Watch what happens when a training video runs as an ad. Completion looks respectable among people who already know you. Cost per result does not move.
Cutroom is shaped for the second viewer. Your own take, cut fast, captions burned in, footage landing exactly where the claim is made.
Two viewers, two completely different builds
| Training video | Feed ad | |
|---|---|---|
| Viewer state | Told to watch it | Trying to scroll past |
| Judged on | Completion and recall | Cost per result |
| Shelf life | Years, updated in place | Turned off in a week |
| Made per month | A handful | Twenty to thirty |
| Best framing | Composed and consistent | Looks like it was not made for you |
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
You direct the cut by marking words, and the machine handles every boundary
Film one take on a phone, up to three minutes. Cutroom transcribes it and drafts the cut. Back comes a 9:16 MP4 with captions burned in.
Highlight a phrase and a clip lands over exactly those words. Swap it if it is wrong, or search millions of free ones.
Delete a line that did not land. The cut rebuilds. Change the pace and the whole piece re-cuts around the new rhythm.
Emphasise the word your claim turns on so it pops in the captions. Correct a mis-heard word and the burned-in text follows.
Write the opening headline yourself or take the drafted one. Turn music on and set how far under the read it sits.
No timeline, no keyframes, no layers. Eleven actions, every one about the ad rather than about software.
Coverage is capped at 40 percent on chill, 45 on normal and 52 on fast. The person stays the spine of the ad, never a voice behind stock.
- Six caption packs, from clean UGC to bold impact, chosen from a menu
- Coverage ceilings of 40, 45 and 52 percent by pace
- Batch is 100 credits and covers transcript, director pass and b-roll search
- You have to be willing to be on camera, or own a photo the avatar can drive
The specification: one take, three minutes, 9:16 MP4
There is no scene library and no template system. Consistency comes from picking the same caption pack, which takes one tap.
The upload ceiling is three minutes. A webinar or a lesson is the wrong source, and a three minute ad is already long.
The output is a 9:16 MP4. That is the shape TikTok, Reels, Shorts and Meta Stories all run.
Nothing of yours is locked onto the export. No logo, no closing frame.
There is no translation. One take, one language, one file.
Those facts are what buys the speed: a finished vertical ad in about four minutes of your attention.
Buy on the viewer, not the feature list, and the answer is obvious
Ask one question. Did the person on the other end choose to watch, or are they trying to get past it?
Chose to watch, permanence matters, buy the training tool. Trying to get past it, volume matters, film forty seconds.
The arithmetic backs the second case. Winners run at 5 to 8 percent of creatives, per Motion's analysis of 550,000+ Meta ads.
At six creatives a month you are drawing under one winner. The fix is more attempts, not a better template.
A second cut of a take already transcribed costs 20 credits a minute. Trying an idea stops being a decision worth debating.
Seven days and 300 credits, no card, is enough for two one-minute ads from fresh takes.
One finished minute from a fresh take is about 120 credits. That number is why a second version becomes a habit instead of a discussion.
Synthesia and Cutroom, judged on a cold viewer
Two rows go to Synthesia and a learning team cannot live without them. The other five are what an ad account is actually buying.
| Synthesia | Cutroom | |
|---|---|---|
| Many languages from one script | Built for global rollout | One take, one language |
| Templates and visual governance | Two hundred videos, one look | Six caption packs, one tap |
| Reads as a real person | Composed by design | Your own take, unpolished |
| Footage on the exact phrase | Scene-based assembly | Highlight, clip lands |
| Thirty variants a month | Not the workflow | 20 CR a minute per export |
| Captions built for a muted feed | Available, not the focus | Burned in on every export |
| Fixing one flat line | Rewrite the scene, re-render | Delete it, the cut rebuilds |
Questions people ask
- Can Cutroom make training or onboarding video?
- It is not built for it. The output is a 9:16 MP4 under three minutes with burned-in captions, which is the right shape for a feed and the wrong shape for a learning system.
- Does Cutroom have avatars at all?
- One. A photo plus a voice-over becomes a lip-synced presenter at 320 credits a minute, with generated voice-over at 70 credits a minute. It exists so a founder who will not film still has a route, not as a roster of characters.
- Why does Cutroom cap b-roll coverage?
- Because past the ceiling the ad turns into stock footage with narration over it, which is the exact thing viewers have learned to skip. The caps are 40 percent on chill, 45 on normal and 52 on fast.
- Is it right for every video?
- It is built for vertical ads aimed at strangers. Learning and internal comms want a training-first tool. Bring the ads here.
Composure earns trust from a captive viewer. Cutroom is built for the other one: your own take, cut fast, captions burned in, footage on the exact claim.