Skip to main content

Rooftop Basketball Dunk

A vertical phone clip of a rooftop dunk, tilting from the sneakers to the rim in one golden-hour beat.

Sports & actionText to video195 credits

Fills the prompt and every setting below into the Seedance 2.5 form. Nothing is submitted until you press generate.

The prompt

A street player drives the baseline, launches off two feet and throws down a one-hand dunk. The camera is a phone held vertical by a friend, tilting up fast from the sneakers to the rim, then loosening as the chain net snaps. Late golden hour on a city roof, long shadows and warm haze between distant towers. Handheld phone look with a faint lens flare, real asphalt and chain texture. Audio: sneaker squeak, chain rattle, one shout.

Settings this prompt was written for

Settings this prompt was written for
FieldValue
ModeText to video
Duration5 seconds
Resolution720p
Aspect ratio9:16
AudioAudio on
Credits195

They are the real price of that exact configuration — computed with the same pricing engine the API charges with, at the resolution, duration and audio setting shown on the card. Change any of those in the generator and the price updates with it.

Why it works

Vertical framing is the whole point of a phone clip, so the prompt fixes the tilt direction and states that the camera is a friend's hand rather than a rig, which keeps the shake human instead of random. The action is a single beat, drive then dunk, with the net snap as the release, so the model does not invent extra dribbles to fill the time. Golden hour is named with its effects, long shadows and warm haze, so the roof reads as a place rather than a grey void.

Mode

Most cases are text-to-video and need nothing but the prompt. The ones marked "needs one reference image" are image-to-video: you supply a photo and the prompt keeps that subject or product locked while the camera moves.

basketball dunk video promptvertical sports clipphone camera dunkgolden hour sports video

Prompts stay in English on every language of this site: that is the language the model was trained on, and translating a prompt changes what it renders. The page around it is translated.