How to write a prompt for video generation on Playground AI
A video prompt needs everything an image prompt does — subject, setting, style — plus one more thing: motion. Without describing movement, the model tends to produce a mostly static clip.
Steps
1. Start with the subject and setting, just like an image prompt (e.g. "a red fox in a snowy forest").
2. Add motion explicitly — what's moving, and how (e.g. "trotting slowly through deep snow, breath visible in the cold air").
3. Add camera movement if you want it — panning, zooming in, static shot, tracking the subject.
4. Add pacing — real-time, slow motion, time-lapse.
5. Keep the scene simple. One clear subject and one clear motion reads better than several moving elements at once.
Vague vs. specific
| Vague | Specific |
|---|---|
| "a busy street" | slow pan across a rainy city street at night, headlights streaking past |
| "waves on a beach" | waves rolling onto shore in slow motion, camera static, golden hour light |
| "a bird flying" | a hawk gliding low over a field, tracking shot, wings barely moving |
Tip
If the motion comes out wrong or too chaotic, simplify the prompt to one moving subject before adding more detail elsewhere.