Better videos

In the Look step the advisor writes one English image prompt per scene. You can edit any of them by hand; edited prompts survive a regeneration.

Pick an image in scene 1. Every later scene gets it as a reference, so face, clothing and colours stay the same.

Live-action, cartoon, isometric, whiteboard or vector icons. Plus your brand kit: logo in every scene, colours in subtitles and the outro.

Every new attempt is a new clip at the price shown. Before you repeat one: sharpen the scene's prompt, try a preview or switch the model - the list shows which model other people kept more often and which one is faster. A job that fails costs nothing.

You can use your own photo instead of a generated image. If people are recognisable you confirm their consent, because the law requires it.