A video prompt is the complete instruction for a generated clip. A motion prompt is the motion portion of that instruction: what moves, how it moves, for how long, and whether the camera follows.
Keeping the two concepts separate helps you diagnose results. If the scene is wrong, revise the video prompt. If the scene is right but movement is wrong, revise the motion prompt.
What belongs in the full video prompt
Google’s video prompt guidance organizes information around subject, context, motion, camera, and style. The full prompt should establish who or what is shown, where the shot happens, what the environment looks like, what the camera does, and the intended visual style.
It should also state constraints that matter to the edit: duration, aspect ratio, continuity with neighboring shots, and elements that must remain stable. This is the creative contract for the whole clip.
What belongs in a motion prompt
A motion prompt narrows the instruction to change over time: direction, speed, force, timing, repetition, interaction, and camera movement. Examples include “the cart rolls slowly left to right,” “steam rises without covering the label,” or “the camera dollies in while the subject stays centered.”
Good motion language is specific and physically plausible. Vague words such as “dynamic,” “cinematic,” or “interesting” describe taste, not movement, and often leave the model to invent action.
How to divide the work
Use the full video prompt to lock subject, scene, style, and output. Then write one primary motion event per clip. Multiple simultaneous actions make causality harder to control and review.
If a storyboard requires several movements, split the idea into cuts. Editing distinct short clips is usually more reliable than asking one generation to perform a complex sequence.
Diagnosing the failed result
If the product, environment, or mood is wrong, do not add more motion words. Clarify subject and context first. If the subject is correct but movement drifts, simplify or slow the action and check whether the source frame supports it.
For continuity failures, compare first and last frames, occlusions, and shadows. A narrower motion request often preserves identity better than stronger style language.
Conclusion
Describe the shot in the full video prompt; describe change inside the shot with a motion prompt. Keeping those jobs separate makes revision faster and generated clips easier to review.
Editorial note: This comparison was drafted with AI assistance, then checked and edited against the sources listed below on September 19, 2026. It does not report an unperformed product test.
Sources
Improve this guide
Report a problem or suggest the next question
Submissions enter a private editorial queue. They are reviewed and deduplicated before any public update or new page is created.