Skip to content

All effectsEpic Dunk

AI Dunk Video: one photo, one-hand slam

An AI dunk video shows you, from one photo, dribbling down a packed pro basketball court and slamming the ball through the hoop with one hand. 6 seconds, vertical 9:16, with the template's sound, from 4 credits.

  • 1 photo
  • 6 seconds
  • Vertical 9:16
  • Sound included

A head-to-toe photo works best: you are seen running from head to shoes.

Video settings

Length6 sEvery video is 6 s, the length of the scene.
Format9:16Every video is 9:16, the frame of the original clip.

This video: 4 credits. Ready in about 5 to 15 minutes; failed generations return your credits. View pricing

Updated

Your video

Your video shows here when it is ready, and stays in My creations.

Real render
  1. Dribble
  2. Drive
  3. Take-off
  4. Slam dunk

The video

What the video shows

A full indoor arena at night, pale hardwood, a referee near the sideline and thousands of people in the stands. Someone in ordinary clothes runs onto the court with a basketball, drives low toward the camera, gathers, takes off and hammers the ball through the rim with one hand while the camera tilts up under the glass backboard. The whole thing lasts six seconds.

That someone is the person in your photo, wearing what they wear in the photo. The court, the camera move and the sound come from a result of the CapCut "epic dunk" template posted by @yyyss7575 on TikTok, which shows an old man in a straw hat making the same play. That post is the reference; it was not made on VideoEffect.

Real render

Second by second

The dunk, second by second

Read from the reference clip, one frame per second.

TimeWhat is on screen
0:00 to 0:02You run in from the left dribbling, the camera tracking at floor height with the crowd behind.
0:02 to 0:03You drop your shoulders and drive straight at the camera and the hoop; a referee stands on the right.
0:03 to 0:04You gather the ball at your waist, plant the last step and jump.
0:04 to 0:05In the air, the ball goes up over your head in one hand, the other arm out for balance, as the camera tilts up.
0:05 to 0:06From under the basket: you slam the ball through the rim one-handed, still airborne.

How it works

How to make an AI dunk video

  1. Upload one photo of yourself. A full-body photo works best.
  2. Choose the resolution. The dunk is always 6 seconds, vertical 9:16.
  3. Press Create and download the MP4. The price is on the button; the file has the sound on it and no watermark.

Price

Six seconds: what it costs

The AI dunk video is the shortest effect in this group: 6 seconds in 9:16 with sound. It costs 4 credits at 480p, 6 at 720p or 12 at 1080p. The smallest pack is 10 credits for $9.99. Pack credits never expire and a failed render gives its credits back. There is no free preview for this effect.

Credits per Epic Dunk video, by length and resolution
Length480p720p1080p
6 seconds4credits6credits12credits

The trend

Where the AI dunk trend comes from

AI dunk videos spread on TikTok in late October and early November 2026 through a CapCut template that its posts caption "Who wants to make an EPIC dunk video just like a pro basketball player?". Posts are collected under #epicdunk and #aidunk, and the trend is searched as AI dunk, epic dunk or the AI basketball dunk trend. Many template results use famous faces; the one this effect follows was chosen because its subject is an unknown old man in a straw hat.

The sound on the clip is the template's own audio from the reference post. The player does not speak or lip-sync.

VideoEffect is not affiliated with CapCut, TikTok, the template's makers or any basketball league or team. No league or team logos are asked for in the video.

Photos

The photo that makes the best dunk

  • Full body, standing. The clip shows you from head to shoes while running, and your shoes and trousers come from the photo.
  • Face visible and well lit. The camera is close during the drive.
  • Wear what you want to dunk in. Jeans and a T-shirt are fine; a jersey only appears if you are wearing one.
  • No hat unless you want one. The straw hat and towel of the reference are not carried over, and a hat appears only if your photo has one.
  • Use your own photo, or one of someone who agreed to be in the video.

The effect was test-rendered once on the same video model on 6 November 2026, from a full-body photo of an AI-made young man at 480p. He hit the dribble, the gather, the take-off and the one-hand slam on the template's timing and kept his face, T-shirt, jeans and sneakers throughout. His T-shirt rode up at the top of the jump, as the shirt does in the reference clip.

Good to know

No three-pointers: the limits

  • It cannot change the play. It is always this drive and this one-hand dunk; no three-pointers, no alley-oops.
  • It cannot put you in a named team's uniform or a named arena.
  • One person only. Nobody else in the arena is replaced.
  • It cannot run longer than 6 seconds or in landscape.
  • A pet can be the subject in principle, because the prompt tells the model to run an animal on its hind legs, but that has not been test-rendered.

Prompt

AI dunk video prompt, in full

Below is an English version of the instructions this effect uses with your photo. They describe the tracking camera and the arena, put the person in the photo in the old man's place, list his straw hat, towel and clothes as things to leave out, and time the five beats of the dunk. It is a translation of instructions written for use with the 6-second reference clip and its sound.

AI dunk video prompt

Follow [Video 1] throughout: copy its framing and camera (a phone held upright, the camera moving with the subject at court level, tracking the dribble from the front and side, tilting up at about 4 seconds and finally looking up at the hoop), its rhythm, and every move, head turn and expression of the subject, as well as the path of the basketball. The scene is the same as in [Video 1]: an indoor professional basketball arena at night with every seat taken, a pale wooden floor, a referee and players in jerseys at the side of the court, a glass backboard and a hoop. Do not add any action or scene element that is not in [Video 1].

Use [Audio 1] as the sound of the whole video, kept exactly as it is: do not add other music, voices or speech.

[Image 1] is the subject's full-body portrait and [Image 2] is the subject's own original photo. There is only one subject: the person in [Image 1]. [Image 2] is used only to see the face clearly: if there are other people in [Image 2], look only at the face of the same person as in [Image 1]; the other people, poses and actions in [Image 2] must not appear in the video. The subject fully replaces the old man in [Video 1] in a big straw hat, with a towel over his shoulder, a beige top with knot buttons and wide black trousers. In every frame the subject is in the old man's place: as close to the camera as he is, as large in the picture, with the same direction of body and head. Weight, build and age all follow the subject's own (as in [Image 1]); do not make the subject older or bulkier. The face follows [Image 2]: the face shape, the features and the size and shape of the eyes are the same as in [Image 2], and do not turn into the face of the old man in [Video 1]; the age looks the same as in [Image 2], and so do the skin texture, the shape of the eyebrows and the lip color; a male subject keeps his own masculine facial lines and a female subject her own feminine facial lines; the expressions are the same as in [Video 1], but no more exaggerated than in [Video 1]. Hairstyle and hair color, build and clothes follow [Image 1]. Whatever is on the face and body is drawn only from [Image 1] and [Image 2]: what is there in [Image 1] and [Image 2] keeps the same look and position and is there from start to finish; what is not there in [Image 1] and [Image 2] is never added. The old man's big straw hat, the towel on his shoulder, his beige knot-button top, wide black trousers and cloth shoes: none of them is drawn on the subject; whether the subject wears a hat, and which clothes and shoes, is drawn entirely from [Image 1] and [Image 2] (no hat in the photos means no hat). If the subject is an animal, it runs on its two hind legs like a person, dribbles with its front paws and dunks with one paw; none of the old man's straw hat, towel or clothes stay in the picture. Ignore the backgrounds of [Image 1] and [Image 2], and leave out any text or captions in [Image 2].

Every move and position of the subject and of the ball is the same as in [Video 1] at the same moment, on the same side (left and right are as seen in the picture).
0 to 1.5 s: the subject runs in from the left of the frame, dribbling, onto the court of a professional basketball arena with every seat taken; the camera runs alongside at court level.
1.5 to 3 s: the subject drops low and drives toward the camera and the hoop, dribbling fast; a referee is on the right of the frame.
3 to 4 s: the subject gathers the ball at the waist with both hands, takes the last step and jumps hard.
4 to 5 s: the subject flies through the air, lifting the ball high overhead with the hand on the left of the frame, the other arm spread out; the camera tilts up.
5 to 6 s: the camera looks up at the hoop; at the rim the subject slams the ball through the hoop with one hand, the other arm spread out, the body still in the air.
The subject does not speak or sing and does not lip-sync; the expression is focused and excited.

9:16 vertical. No split screen; the subject is the only one dunking in the picture, and the spectators, the referee and the players at the side stay where they are in [Video 1]; no captions, text, watermarks, logos or stickers on the picture, and no league or team logos or names anywhere in the arena. The subject's face stays consistent from start to finish, the same as in [Image 2]; hairstyle and clothes do not change either, the same as in [Image 1]. The old man's big straw hat and the towel on his shoulder never appear.

FAQ

Questions and answers

Something else? Contact us or see pricing.

What is an AI dunk video?

An AI dunk video is a short AI-made clip of a person dunking a basketball in a packed arena, made from a photo. The person never touches a ball; the play comes from a template clip.

How do I make an AI dunk video of myself?

Upload one photo on this page, pick a resolution and press Create. Six seconds of video with sound come back as an MP4.

Is the AI dunk video free?

No. It costs credits, from 4. There is no free preview, and failed renders refund their credits.

Is this the CapCut epic dunk template?

It follows a result of that template, so the play and the arena are the same. You do not need CapCut to use this page.

Can I wear my own jersey?

Yes, if you are wearing it in the photo. Clothes are taken from the photo.

Which arena or team is in the video?

None by name. The setting is a generic pro arena, and the prompt asks for no league or team logos.

Can I dunk with a friend in the same video?

No. The effect replaces one player.

Can my dog dunk?

The prompt has a line for animal subjects, but only a human photo has been test-rendered.

How long is the video?

Six seconds, vertical, the length of the template clip it follows.

Epic Dunk

Make yours

The price is on the Create button before anything is charged, and a render that fails gives its credits back.