Turn one prompt into a full sequence — scene by scene, consistent characters, native audio — then refine it in the editor.