Motion from a still
Describe motion without accidentally redesigning the source image
Image-to-video prompts should explain what moves, how it moves, what the camera does, and what must remain untouched. The workflow keeps those instructions distinct.
Protect the starting frame
List the subject, layout, colors, wardrobe, product details, text, and background elements that must not change.
Add motion in layers
Separate primary subject action, secondary environmental motion, camera movement, and timing so instructions remain readable.
Prevent unwanted invention
Use targeted negative instructions against new objects, altered branding, warped anatomy, abrupt camera jumps, and composition drift.
Quick answers
Questions about this workflow
Should I repeat everything in the image?
No. Lock the important details and focus the rest of the prompt on motion.
Can it plan lip-sync?
It can plan dialogue cues, but actual lip-sync quality depends on the video model.