Rooftop chase action sequence
This prompt builds a short action sequence from three shots, each with one clear move: a tracking run, a slow-motion leap and a landing close-up. Keep the hero's outfit the same in every shot so she stays consistent.
Action sequence, 16:9, 10 seconds, three shots. Shot 1: shaky tracking shot behind [HERO, e.g. a young woman in a black jacket] sprinting across [SETTING, e.g. rooftops in old Delhi at dusk]. Shot 2: low-angle side shot as she leaps the gap between two buildings, slow motion at the peak. Shot 3: frontal close-up as she lands, rolls and looks back, breathing hard. Dust, washing lines, orange sky, high tension, no weapons. No on-screen text.
Action reads best when each shot has one clear move. Keep the same outfit description in every shot so the hero stays consistent.
From copy to clip in four steps
- 1
Open Higgsfield.
- 2
Paste the prompt and fill in the [BRACKETED] fields. Add a starting image if the tool supports it.
- 3
Set the aspect ratio and length in the tool.
- 4
Generate two or three takes and adjust one detail at a time.
About this prompt
Which AI tools work with this prompt?
It is written for Higgsfield and Seedance. Paste it into your preferred tool and adjust the wording if the result is off.
What do I need to fill in?
Replace these fields before you send it: HERO, SETTING. Each field shows an example in the prompt.
Why does the result differ between video models?
Each model interprets camera moves, motion and audio differently. Keep the prompt, regenerate a few times and change one detail at a time.
