Understanding Generation Prompts
How your workflow settings are translated into the prompt sent to the AI model.
You never have to write a prompt in ScriptMotion. The workflow builder collects structured choices, and the platform composes the prompt from them at generation time.
What goes into the prompt
Each workflow step contributes a different part:
- Format & Layout sets the shot style, aspect ratio, and duration, and determines which model the prompt is written for.
- Character & Product contributes the subject description and any reference images the chosen model can use.
- Scenes and References contributes the setting — as a description, and as reference images or a background image on models that accept them.
- Concept & Script contributes the narrative and any spoken lines.
- Audio contributes the delivery direction: tone, pacing, and ambience.
Brand-kit enforcement is opt-in, not automatic. Turn it on per video with the Brand Kit toggle in Concept & Script (Step 4) — off by default. When it's on, the brand kit's tone and visual direction are threaded into that video's prompt; when it's off, the video generates from your workflow choices alone with no brand-kit influence.
Prompts are composed fresh at generation time from your current settings, so changing a step and regenerating always reflects the change. There is no stale, saved prompt to clear.
Getting better results
- Be specific in the script rather than the styling. The layout and model already carry the visual style; the script is where detail pays off.
- Give the model references it can actually use. Check the Image inputs column in AI Models Explained — a product shot attached to a model that ignores product references will not change the output.
- Change one thing at a time. Regenerating after a single edit tells you what that edit did; changing four things at once does not.