Comparison
Runway makes shots. Cutroom makes the thing you upload.
Generative models produce shots nobody could film at any budget. Runway is the strongest of them. A folder of striking shots is still not an ad, and assembling it is still an evening. Here is where the second half starts, what a finished ad costs, and how Cutroom builds one from a take.
- frames of video Cutroom generates
- 0
- one batch, one drafted cut
- 100 CR
- coverage ceiling by pace
- 40-52%
For a shot that cannot exist, a generative model is the only supplier
A product dissolving into its ingredients. A camera move through a wall. A location nobody is flying to.
That is a real capability with no substitute. For concept work and film it changes what a small team can attempt.
It also fixes the coverage gap when a stock library has nothing close to the specific thing you need on screen.
It is the fastest route to a look nobody else has. Stock libraries are shared, and a generated shot is not.
For a brand piece where the imagery is the point, that is the entire reason to buy it.
Here is what a shot does not include. Something still has to order them, time them, caption them and give them a spine.
In a direct-response ad the spine is a person making a claim. Shots support a claim, they do not replace it.
Cutroom starts at the spine. One take of up to three minutes goes in.
Back comes a finished 9:16 MP4, captions burned in and coverage sitting on the words you marked.
A folder of beautiful shots is not an ad, and assembling it is still an evening
Without a spine, the output is a montage with narration on top. That is the format a feed skips fastest.
There is a second problem specific to generation. The showreel you saw is a survivor, not a sample.
Getting a usable clip takes attempts. Every attempt costs money and time before assembly starts.
None of that effort produces captions, a hook, or a reason for a stranger to keep watching.
Here the cut arrives drafted. Coverage is already placed and captions are already burned in.
You adjust it by marking words. Eleven marks, then an export at 20 credits a finished minute.
The two sit next to each other rather than against each other. One supplies shots, one supplies the ad.
Most accounts that buy both use the generated shot as coverage over the phrase it illustrates.
The two halves of making a video ad
Generative tools supply the left half. Cutroom supplies the right half, from a 100-credit batch.
Shots exist
Filmed, stock, or generated
A spoken claim
Cutroom needs this, up to 3 min
Transcribed and drafted
Batch: 100 CR
Marks on the transcript
Highlight, delete, pace
Captioned 9:16 MP4
20 CR a minute
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
Highlight a phrase and coverage lands there, generated or not
Upload a talking-head take of up to three minutes. It comes back transcribed, drafted and captioned as a vertical MP4.
Highlight a phrase and a clip lands over exactly those words. Swap it, or search millions of free ones.
Delete a line and the cut rebuilds around the gap. Change the pace and the whole piece re-cuts.
Emphasise the word the claim turns on so it pops in the captions. Fix a mis-heard word and the text follows.
Coverage is capped by pace: 40 percent on chill, 45 on normal, 52 on fast. The person stays the spine of the ad.
There is no timeline, no keyframes and no compositing surface. This is a cutting room, not a post house.
Trim the opening or the ending, write the headline over the first three seconds, and choose how each clip enters.
- Six caption packs, adjustable for colour, weight, size and position
- Five entrances for a clip: cut, whip, punch, glitch or sweep
- Music from a described style at 20 credits a track, sitting under the read
- Export 20 credits per finished minute, from a take already batched
A spoken take in, an uploadable file out: the specification
One talking-head take of up to three minutes goes in. One 9:16 MP4 comes out, captions burned in.
There is no text-to-video and no image-to-video. If nobody speaks, there is nothing here to work with.
There is no masking, no keying, no motion graphics and no colour grading.
There is no shot library. Coverage comes from a stock search at the moment you highlight a phrase.
The avatar module is the only synthetic thing here. A photo you own plus a voice-over, at 320 and 70 credits a minute.
There is no timeline underneath and nothing of yours gets stamped on the finished file.
Those are the walls. The number they do not touch is finished ads per month, which is next.
Thirty creatives a month is an assembly problem, not an imagery problem
Winners run at 5 to 8 percent of creatives, per Motion's analysis of 550,000+ Meta ads. The count decides the account.
A second cut of a take already transcribed costs 20 credits a minute and about four minutes of attention.
That is what turns an uncertain angle into an attempt rather than a note nobody returns to.
One finished minute from a fresh take is about 120 credits, which is roughly what a handful of generation attempts costs in time alone.
Seven days and 300 credits with no card is two finished ads and a straight answer about your bottleneck.
Keep the generative tool for the striking piece of imagery a campaign genuinely needs.
Bring the claim here and let the coverage land on the phrase it illustrates.
Runway and Cutroom, row by row
The top two rows go to Runway, and an image that cannot exist should weigh them heavily. The rest is the half after the shot.
| Runway | Cutroom | |
|---|---|---|
| Generate a shot from a prompt | Text or image in | Nothing is generated |
| Compositing, effects, any ratio | A real post toolkit | 9:16, captions, coverage |
| Captions burned in for a muted feed | Shots have no words | Six packs, one tap |
| Cut rebuilds when a line goes | Reassemble by hand | Delete, and it recompiles |
| A finished ad from one take | Shots, not structure | Drafted cut, captioned |
| Coverage tied to a phrase | Clip by clip | Highlight, clip lands |
| Cost of the twentieth ad | Attempts, then assembly | 20 CR and four minutes |
Questions people ask
- Can I use generated clips inside Cutroom?
- Not as an upload. The source is your talking-head take, and coverage comes from the stock search inside the product. If a specific generated shot is central to the ad, assemble that piece somewhere with a timeline.
- Does Cutroom generate any video at all?
- Only through the avatar module, which lip-syncs a photo you own to a voice-over at 320 credits a minute. There is no scene generation, no background generation and no text-to-video route.
- Why cap b-roll coverage at all?
- Because above the cap the ad stops being a person and becomes footage with narration, which is the exact thing viewers skip. The ceilings are 40 percent on chill, 45 on normal and 52 on fast.
- Who should not buy Cutroom?
- Anyone whose value is visual invention and whose deliverable is the shot itself. Buy the generative model for that. Come here when the shot has to sit inside an ad somebody believes.
Striking shots with no claim inside them are a mood board with a render bill. Cutroom returns the finished 9:16 ad, captions burned in and coverage on the claim.