Prompt library
The point of a good prompt is not to sound clever. The point is to help the model lock onto one scene, one motion pattern, and one visual direction quickly enough that your first draft is useful.
The prompt set below is built for practical iteration: product work, short ads, stylized scenes, and image-to-video workflows where you want clear control instead of generic prompt sludge.
Subject + action + environment + camera + lighting is usually a better starting point than long style dumps.
Start with these templates, then swap only one variable per iteration so you can see what actually changed the output.
If you already have a strong first frame, move to image-to-video instead of forcing too much control into text alone.
Landing pages, product launches, paid social
This works because it keeps the subject count low and gives the model a clear camera and lighting instruction.
Brand mood pieces, vertical social clips
Use this when you want motion and attitude without turning the scene into a noisy crowd shot.
Restaurant promos, short ads, creator content
Macro prompts perform better when the action is singular and the lighting direction is obvious.
Character intros, stylized fan edits, motion tests
This pattern is better than vague action prompts because it defines the subject, the entry move, and the environment.
Tourism, real estate, cinematic openers
Landscape prompts improve when you specify both camera movement and the time of day.
Performance ads, creator campaigns, app promos
This keeps the scene believable and avoids the over-polished look that often hurts ad-style outputs.
SaaS launches, dashboard teasers, demo clips
Use image-to-video if you already have interface mockups. It gives you tighter control than a fully generative prompt.
Character tests, storyboards, repeatable scenes
If consistency matters, reduce scene complexity and keep the motion narrow instead of asking for too many changes at once.
These templates match common search intent and work best as first-draft starting points.
Use this when you need a controlled ad-style first draft with one clear product subject.
Best used in image-to-video mode when your first frame already matches the campaign style.
Keep camera direction and time-of-day explicit to reduce inconsistent landscape motion.
Food prompts perform better with one singular action and clear lighting language.
Good starter template for users searching Seedance with spelling variations like "Sea Dance".
Use this as a first-pass dance prompt, then switch one variable at a time (camera, tempo, or background) to keep motion stable.
Best for image-to-video when pose consistency matters more than scene complexity.
Use this as a stylized scene test with one human action and one secondary subject to avoid motion clutter.
Open the generator, test one prompt with one mode, then compare it against a stronger reference image or a tighter version of the same prompt. That is usually where the real gains show up.