Direct the ad by marking words, not by dragging clips

Upload one talking-head take of up to three minutes. It comes back transcribed with word-level timing. Cut, covered with footage, captioned and headlined, before you touch anything.

Then the transcript is handed to you as a surface with a fixed set of moves on it. There is no cursor. There is nothing to type into.

Each move is a directing decision rather than an editing chore. Highlight a phrase and footage lands over exactly those words. Delete a line and the cut rebuilds around the gap.

That is the whole vocabulary and it is short on purpose. What comes out is a 9:16 MP4 with the captions rendered into the picture.

The standard version of this idea works one way. The tool transcribes your footage, shows you the words, and lets you delete text. Delete the text, delete the video.

That is a better interface than dragging clip handles. It changed how a lot of people work. Here is the part nobody says out loud.

Once the transcript is a document, you treat it like a document. You tidy a sentence. You reorder a paragraph because it reads better. Forty minutes later you have a beautifully edited read, and the read was never the deliverable.

The transcript is not useful because it is text. It is useful because it is the only view of your footage where your argument is visible in one glance.

So every move here changes the ad. There is no move that only changes the document.

The moves are a known list, so each one applies to your original take rather than to the last version of the edit. Your changes are a short set of instructions. The cut is recomputed from them every time.

So captions do not slide because you removed a line above them. Footage does not hang over the wrong sentence. There is no accumulated state to corrupt.

An open text field could not offer that. If you can write anything, the tool has to guess what you meant. A tool that guesses right nine times in ten is a tool you check every time.

Checking every time is the cost nobody prices in. It turns an eight minute change into a twenty minute change, silently, on every pass.

Put a number on the difference. A nine word change to your hook is eight minutes here. The second nine word change is eight minutes again. That second number decides how many hooks you ever test.

A timeline is better at two jobs and both are worth naming. It trims to the exact frame. It assembles any number of sources into one edit.

If either of those is the job in front of you, that is the tool for it. Neither one is what a vertical ad needs.

A vertical ad is one person, one read, and eight more versions of it by Friday. Nothing on a timeline was designed for the eighth version.

That is where a closed move set wins, and it wins by a distance. The first cut arrives finished and every version after it is minutes of work.

The scope here is one talking-head take of up to three minutes returning one 9:16 ad. Removal works by the line, because yanking a word out of spoken audio leaves a seam you can hear.

The cut follows your read in the order you spoke it. Sentences joined out of order carry mismatched breath, and a viewer hears the join before they hear the claim.

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

Changing the pace, the captions, the headline or the clips costs nothing. You are charged when you download.

Put a number on it. A batch plus a thirty second export is 110 credits. Basic's 2,500 credits a month is around twenty-two finished ads, made and downloaded.

The trial is 7 days and 300 credits with no card, which is two finished short ads with change left. Basic is $39.99 a month for 2,500 credits. Premium is $79.99 for 5,000. A $15 top-up adds 1,000.

Set that against the document-editing evening. Not in money, but in the ads that never got made because the second attempt cost the same as the first.