How to make a particular kind of video, and what to choose before you generate. Each guide opens with a video made by the app itself — including the ones showing the real interface.
Pick 9:16 before anything is generated, and every scene is framed for it — with captions lifted clear of the platform's own buttons.
Paste the words you want said. They are narrated verbatim, one scene per sentence, with captions timed to the voice.
Write the steps, watch the cut, and drop any scene that drags — using your own screenshots, held still so the interface stays readable.
One episode per run, each knowing it is part N of M, with the schedule stopping itself after the last part.
Connect the app once, then ask in chat. Claude can create a video, list its scenes, drop one and re-cut — without you opening an editor.