Skip to main content
Once the story is written, the app stops on “Review the story before paying for images”. You see every shot, its line, its length, and the cost of what comes next. It is the cheapest moment to fix things: nothing has been drawn yet.

​
Three levels of correction

You can also change “Video length”: “The story is rewritten to this length, and the video price changes with it.”

​
Writing a good instruction

An effective instruction points to the place and says the wanted result, not the vague problem.
Other examples suggested by the app:
Several precise corrections in a single instruction are better than several successive rewrites: every rewrite costs credits.

​
The virality analysis

Before approving, you can run “Analyze virality”: a score out of 100 across 8 criteria, and the three fixes that earn the most points. “Apply” rewrites the part with the fix as the instruction (it costs credits, like “Change the story”). The criteria are detailed in The rules of a good prompt.

​
After the video

Once the video is edited, each scene can still be corrected without redoing everything:
  • “Rewrite the scene”: “Change what happens or what is said”.
  • “Redo the image”: say what is wrong, for example “He looks at the camera, more light on his face”. Only that scene is redone.
  • “Redo the clip”: keeps the image and the voice, or goes back to an older version.
  • “Animate the scene” or “Switch to slow zoom”: an animated clip or a slow zoom on the image.
  • “Add a scene”: a new scene, written and animated, wherever you choose.
To go further (replacing a music track, regenerating a voice), see Editing.

​
Common mistakes

  • Approving too fast. After “Approve the story”, the full video is launched with that breakdown as is. Review at least the first line and the ending.
  • Fixing an image by rewriting the story. If only the visual of a scene is wrong, “Redo the image” is enough.
  • An instruction without a shot number. “She must be angrier” on a story of fifteen shots leaves the AI guessing where.