Upload one take. A finished 9:16 ad comes back

The upload is one file. A person talking to a camera. Up to three minutes, filmed on a phone, in a normal room.

The machine transcribes it with word-level timing. It decides where the cuts fall. It searches for footage and cuts each clip to the run of words it covers. It writes captions timed to your speech and drafts an opening headline.

What downloads is a 9:16 MP4 with the captions rendered into the picture. Nothing is left to assemble. Nothing needs a second tool.

One trip through the machine is 100 credits. Export is 20 credits per minute of finished video. A thirty second ad costs about 10 to download. Call the whole thing 110 credits.

The 7 day trial gives you 300 credits and asks for no card. That is two ads made and downloaded, with change left over.

There is no timeline in the editor. Nothing to scrub. No clip handles. No caption timing done by hand at eleven at night.

You read your own words and mark them. Highlight the phrase where you name the product. A clip lands over exactly those words, starting on the first and gone on the last.

Delete the line where you wandered. The cut rebuilds around the gap. Set the pace to fast and the whole read is re-cut tighter.

That is the mechanism, and it matters for one reason. An instruction that points at a word cannot be moved by an edit above it. Nothing drifts when you change your mind.

So nothing in that list is a repair job. You never re-time a caption because you deleted a sentence above it. You never nudge a clip back onto the word it slipped off. Those two chores are most of what a timeline evening is made of, and they are why attempt seven never gets built.

Motion looked at more than 550,000 Meta ads. Roughly 5 to 8 percent of creatives become winners. Call it one in thirteen to one in twenty.

Now count what your account shipped last month. If the answer is six, the explanation is arithmetic rather than taste.

An under-sampled account looks identical to an uncreative team from the outside. It gets treated as one in the quarterly review.

So the question that decides a quarter is not how good attempt one is. It is what attempt seven costs once you have paid for six.

Here attempt seven costs 10 credits and eight minutes. The batch was paid for. Only the export is charged again.

One take becomes one ad. The uploader accepts up to three minutes of it, which is longer than any vertical ad should run.

Somebody points a camera at their own face and talks. That take is the asset. It is also why the finished ad sounds like a person rather than a script.

Removal works by the line. Delete a whole line, or change the pace and let the read tighten. Pulling one word out of spoken audio leaves a join you can hear.

Two recordings cannot be joined into one ad. If you filmed the hook separately from the body, run each as its own test and let the numbers choose between them.

The export is a 9:16 MP4. That is the shape a vertical feed plays, and the composition is built for it from the first frame.

Audio decides more than people expect. A quiet carpeted room with a phone on a stand beats a good camera in a kitchen. Bad audio makes the transcript worse, and the transcript drives the footage.

A batch is 100 credits. It covers transcription, the directing pass and the footage search. Export is 20 credits per minute of finished video.

Marking the transcript is free. So is restyling the captions, swapping clips and rewriting the headline. Only the download is charged.

So eleven exports of a thirty second ad is 110 credits. That is less than one more trip through the machine. It is the whole pricing argument.

You start with 7 days and 300 credits, no card. Basic is $39.99 a month for 2,500 credits. Premium is $79.99 for 5,000. A $15 top-up adds 1,000 credits when a month runs long.

Basic works out at about twenty finished short ads a month. Set that against six, and against a hit rate of one in thirteen to one in twenty.