Genjutsu AI

2 photos · the 15 s kitchen clip · 9:16 with sound

Wipe Me Down Dance AI: Template, Prompt or Your Own Clip

The Wipe Me Down dance is the low, loose bounce: lean toward the camera, sink into a wide crouch and flick both hands at the floor on every beat, set to “Wipe Me Down”, the 2007 Foxx single with Lil Boosie and Webbie. The 2026 AI trend recasts one 15-second kitchen clip in which a dancer goes first and a friend takes over. It is not the 2020 “Wipe It Down” mirror challenge. Upload two photos, or the same photo twice, and Genjutsu AI follows that clip’s moves, camera and sound, from 8 credits.

  • Two photos, or one photo twice
  • 5, 10 or 15 s · 9:16 with sound
  • From 8 credits
  • Failed jobs refund credits

Photo 1 · Dances first

Photo 2 · Takes over at 9 s

One person per photo, face clear. Dancing alone? Put the same photo in both slots.

JPG, PNG or WebP · up to 30 MB · 300–6000 pixels on each side. Each dancer wears the clothes from their own photo.

Resolution
Video length

15 seconds is the whole clip. 5 seconds shows only the first dancer; 10 seconds ends as the friend jumps in.

15s · 480p · vertical 9:16 MP4 · the clip's sound

Costs 117 credits

Seedance 2.5 checks both photos and the reference clip before it starts. A rejected or failed job returns its credits.

Your uploaded photos are private and used only for your generation. Privacy policy

The 15-second kitchen clip the AI trend copies. Photo 1 replaces the first dancer, Photo 2 the friend; the moves, camera and sound stay.

How to make a Wipe Me Down dance AI video from two photos

  1. 01

    Pick who dances first

    Photo 1 leans into the camera and dances first; Photo 2 waits by the counter and takes over at about 9 seconds. Use one clear photo per person, face visible, ideally waist-up or full length. Dancing alone? Put the same photo in both slots.

  2. 02

    Choose the length, model and resolution

    Keep 15 seconds for the whole routine with the hand-over. 5 seconds shows only the first dancer and 10 seconds stops as the friend jumps in. The price on the button updates with every change.

  3. 03

    Generate

    Confirm that everyone in the photos agreed, then start. On Seedance 2.5 both photos and the clip are reviewed first; a rejected or failed job returns its credits. You can leave the page and come back to the same task.

  4. 04

    Download the MP4 and post it

    The finished file is a vertical 9:16 MP4 with the clip’s sound. Upload that file directly to TikTok, Reels or Shorts instead of a screen recording, and label it as AI.

The generator above follows one clip: the 15-second kitchen take the 2026 AI trend copies. Photo 1 leans into the phone, drops into the bounce and does a small tucked jump; Photo 2 waits by the counter and takes over at about the 9-second mark. Each dancer wears the clothes from their own photo. Put the same photo in both slots and you dance both parts.

What the Wipe Me Down dance AI trend copies

Most Wipe Me Down dance AI videos rebuild the same short scene, so it helps to know which song, dance and clip the name points to.

The song

“Wipe Me Down”, the 2007 Foxx single with Lil Boosie (now Boosie Badazz) and Webbie, from Trill Entertainment’s Survival of the Fittest compilation. The Wipe Me Down dance song reached No. 38 on the Billboard Hot 100 (

Wikipedia

).

The dance

Lean toward the camera, sink into a wide, low crouch and bounce side to side while both hands flick toward the floor on each beat, as if wiping it.

The dancer behind the revival

Carnell Nichols Jr., whose compilations and the Team Carnell TikTok account present him as the original concept creator; the masked “Louisiana Michael Myers” takes tag him.

The clip AI versions copy

A 15-second handheld kitchen video by @dadxander (August 2026): one dancer goes first, a friend takes over at about 9 seconds.

Not the same as

The 2020 “Wipe It Down” challenge, where people wipe a mirror and appear in a new outfit to BMW Kenny’s song (

TIME

).

The Wipe Me Down dance trend lives on TikTok as a meme as much as a dance: the funny takes come from who does it and where, not from new choreography. That is why the AI version works with any two people who agreed to it, and why so many people search for the same clip.

Template, prompt or your own footage: which route fits

There are four ways to make the video, and they fail in different ways. Choose by what you already have.

RouteYou supplyWhere the moves come fromWhat usually goes wrong
This page’s templateTwo photos, or one photo twice

The original 15-second kitchen clip, frame by frame, with its sound

A face that drifts in the low crouch when the photo is small or side-on

An edit template in a video appWhatever photos the template asks for

Depends on the template: some only add effects, transitions and the song around your still photo

A slideshow look; length, charges and watermarks depend on the app and plan

One photo and a promptA photo and a written descriptionThe model invents its own room, beat and movesMoves that only loosely resemble the dance, frozen poses, no song
Your own recording, recastYour clip and the photo of who should appear in itYour footage

Needs a clip where the dancer is fully in frame; use motion transfer

  • You want the viral version with a friend, partner or yourself twice: use the template on this page.
  • You want a different room, a pet or a style the clip does not have: use one of the prompts below in an image-to-video tool, knowing the moves will be its own guess.
  • You already filmed yourself doing the dance: keep your moves and change who appears with Genjutsu motion transfer, or swap two people in your clip with duo swap.

Wipe Me Down dance AI examples

Four takes on the Wipe Me Down dance video: an AI recast of the kitchen clip, a prompt-only AI version, a wedding and an animal-inspired one.

Two new characters, same kitchen routine

The kitchen clip recast with two costumed characters: one drops low first, the other takes over. This is what the two-photo route changes.

One photo plus a prompt

A one-photo AI version made without a reference clip: the model invents its own room, beat and moves around the low lean.

At a wedding reception

A member of the bridal party takes the dance floor to the original 2007 song at a wedding reception.

The animal-inspired version

A creator copies a flamingo’s moves on the trend’s 15-second sound, cut against footage of the bird.

Notice the difference between the first two: the recast keeps the original timing and hand-over, while the prompt version moves on its own beat in a room the model chose.

Make it from two photos, or one photo twice

SlotRole in the clipOn screenPick a photo where
Photo 1Leans into the lens, drops low, jumps, wipes the floor0 to about 9 s, then watches from the side

The face is large and sharp: the opening is a close-up into the camera

Photo 2Waits by the counter, jumps in and takes overIn the background at first, centre stage from about 9 s

The whole body shows, so trousers and shoes are not left to the model

Same photo in bothYou dance both parts, as two people with the same faceAll 15 sYou are facing the camera with your arms visible

All three lengths start at 0 seconds. 5 seconds is only the first dancer’s drop, 10 seconds ends as the friend jumps in, and 15 seconds is the whole routine including the hand-over. If you are making a couple or best-friend version, choose 15 seconds or Photo 2 barely appears.

Copy-ready Wipe Me Down AI prompts

You do not need a prompt for the template above. These are for image-to-video tools without a reference clip, and for versions the kitchen clip cannot do, such as a pet. Paste the prompt, attach the photo or photos it mentions, and choose vertical 9:16 if the tool asks.

Solo portrait that keeps your face
Use my photo as the only reference for the dancer's face, hair, skin tone, build and clothes. Full-body shot, phone camera at chest height, in a small kitchen at night.
The dancer leans toward the lens with a grin, hops back and drops into a wide, low crouch with knees bent and back tilted forward, rocking side to side while both hands flick down toward the floor on every beat, as if wiping it. A small tucked jump in the middle, then back down into the low bounce.
Keep the face identical to the photo in every frame, including the low crouch. One person, natural hands and feet, real-time speed, vertical 9:16, no captions or watermark.

Upload a sharp, front-facing photo. If the tool has a strength or motion setting, keep the face reference high and the motion medium.

Couple or best friends, one after the other
Two people from my two reference images, Person A and Person B, in one continuous handheld vertical shot in a living room.
Person A dances first: leans into the camera, drops into a wide low bounce with both hands flicking at the floor on each beat. At the halfway point Person A steps aside laughing and Person B jumps in and does the same low, loose bounce while Person A watches with arms folded.
Each person keeps the face, hair and clothes of their own image the whole time; never swap or blend them. Exactly two people, natural hands, no text or watermark.

Attach the first person’s photo first. Tools that accept only one image will merge the two people; use the template on this page instead.

Your dog doing the bounce
Use my dog's photo as the only reference for the dog: same breed, coat colour, markings and size. Vertical phone video, kitchen floor, camera low.
The dog stands on all four legs, lowers its chest into a playful bow and bounces side to side in rhythm, front paws tapping the floor on each beat, tail wagging; it stays a real dog with a dog's body and moves the whole time.
Natural fur and paws, no human hands, no clothes, real-time speed, no captions or watermark.

Keep it a real dog on four legs; asking a dog to crouch like a person is what produces extra limbs.

Two people from an old family photo
Two people from my two reference images, scanned from an old family photo: keep their faces, ages, hairstyles and clothes, but show them as present-day colour video with clean, sharp detail.
In a family kitchen they take turns doing the Wipe Me Down dance: lean toward the camera, sink into a wide low bounce and flick both hands toward the floor on every beat; the second person joins in laughing at the halfway point.
Two people only, natural hands and feet, vertical 9:16, no text or watermark.

Crop each face from the scan into its own image first, as large as the scan allows.

Negative prompt (where the tool has one)
different face, face changes when crouching, extra fingers, merged hands, extra legs, frozen pose, slideshow, slow motion, morphing body, duplicated person, blurry face, captions, watermark

For more patterns, see the Genjutsu AI prompt guide.

Couples, friends, family photos and pets: what works

You wantOn this pageOtherwise
You and your partner or a friend

One photo each. Whoever should open the video goes in Photo 1; the person who “wins” the dance goes in Photo 2.

The couple prompt, if the tool accepts two images
Just youThe same photo in both slots: you dance first and then take overThe solo prompt
Grandparents from an old print

Scan or photograph the print, crop the two faces into separate photos, at least 300 px on the short side

The family prompt
Three or more peopleNot possible: the clip has two dancers

Record or find a clip with that many people and recast it in the main editor, up to nine photos

A dog or catNot supported: both roles are people

The pet prompt, or your own clip recast with your pet’s photo in the main editor

Use photos of adults who agreed to be in the video, and leave strangers out of the frame.

Cost per video and per batch, free credits and watermark

Credits below are per video, including the reference processing, not per second. The button shows the same number before you start. The default, Seedance 2.5 at 15 seconds and 480p, is 117 credits; the cheapest full clip is Mini at 480p for 23 credits; the most detailed, Seedance 2.5 at 1080p, is 464.

Wipe Me Down dance AI credits per video, by model, resolution and length
Model5s10s15s
Seedance 2.5 · 480p3978117
Seedance 2.5 · 720p88175262
Seedance 2.5 · 1080p155310464
Seedance 2.0 Mini · 480p81523
Seedance 2.0 Mini · 720p163248
Seedance 2.0 Fast · 480p163248
Seedance 2.0 Fast · 720p3569104
MiniMax H3 · 720p3977116
Seedance 2.0 · 480p275481
Seedance 2.0 · 720p58116174
Seedance 2.0 · 1080p146291436

Planning a series, or a campaign with several people? The batch table multiplies the full 15-second clip. Dollar figures use the Pro monthly plan’s credit value ($29 for 780 credits); they are not separate checkout prices.

Credits for 1, 10 and 100 full Wipe Me Down clips
Full 15 s clip on1 video10 videos100 videos100 videos ≈ USD
Seedance 2.0 Mini · 480p232302,300$85.51
Seedance 2.0 Mini · 720p484804,800$178.46
MiniMax H3 · 720p1161,16011,600$431.28
Seedance 2.5 · 480p1171,17011,700$435.00
Seedance 2.5 · 720p2622,62026,200$974.10
Seedance 2.5 · 1080p4644,64046,400$1,725.13
Free allowanceNew accounts receive 3 credits, valid for 7 days. The cheapest video costs 8, so you need more credits before generating. Watching the examples costs nothing.
Smallest purchase$5 Trial Pack gives 24 credits. That covers one full 15-second Mini video at 480p (23 credits).
WatermarkUntil your first purchase, videos carry a moving genjutsu-ai.app watermark. Any purchase unlocks the clean original of the videos you already made and of every new one.
Failed or rejected

The job’s credits come back automatically, including when Seedance 2.5’s review rejects a photo.

A finished video you do not likeAnother attempt is a new generation and uses credits again.

See plans and credit packs to buy credits.

Why it looks like a slideshow, crops your body or changes your face

Check the photo, the downloaded file and the posted version separately: a clean upload cannot repair a face that was already wrong in the MP4.

What you seeUsual causeWhat to do
A still photo with a zoom, or a slideshow

The tool animated the photo without a motion reference, or the prompt described a pose instead of movement

Use a reference clip (the template here), or describe continuous motion at real-time speed

The face changes in the low crouch

A small, side-on, filtered or sunglasses photo gives too little to hold on to when the head tilts down

A sharp, front-facing photo without heavy filters; try 720p or 1080p for more facial detail

Hands melt into the floor, extra fingers or legs

The floor wipe is fast and the hands cross the body, so the model has to guess them

A photo where both hands and legs are visible; another attempt may fix it, without a guarantee

Head or feet cut offA tight crop in the photo, or a video frame saved too close

A waist-up or full-length photo with space above the head; the output is always 9:16

The second dancer looks like the first

Both photos show similar people or outfits, or the same photo twice (which is intended)

Two different, clearly lit photos; the order decides who goes first

Sharp download, blurry postScreen recording, repeated re-exports or chat-app compressionUpload the downloaded MP4 directly to the app

Photos, privacy and the song before you post

  • Consent. Use your own photo or photos of adults who said yes. A friend’s face in a dance meme is still their face.
  • Privacy. Uploaded photos are stored privately with your account, sent to the video model only as short-lived signed links, used only for your generation and never shown on this site.
  • The song. The download carries the clip’s audio, the trend’s 15-second cut of “Wipe Me Down”. Posting it on TikTok or Reels works the way any trend sound does there; a brand ad needs a music licence, and buying credits does not grant one.
  • Label it. Mark the video as AI-generated when you post it, so nobody thinks the people in it really danced in that kitchen.

Want a different dance? Try the Toc Toc AI dance template.

Last updated: 2026-10-23

Wipe Me Down dance AI: questions

What prompt keeps my face realistic while my portrait does the Wipe Me Down dance?

Say that the photo is the only reference for the face, hair, skin tone and clothes, ask for a full-body shot at chest height, describe the moves (lean in, wide low crouch, both hands flicking at the floor on every beat) and add “keep the face identical in every frame, including the low crouch”. The copy-ready version is in the prompts section. With a prompt alone the model still invents the moves; on this page the reference clip supplies them and you write no prompt at all.

Can my dog do the Wipe Me Down dance?

Not with this template: it replaces two people and expects a person in each photo. For a dog, use the pet prompt on this page in an image-to-video tool, or record your own clip and recast it in the main editor’s Character mode with your dog’s photo, where an animal keeps its own body. Four legs move very differently from a crouching person, so expect more drift than with people.

Can I turn an old family photo into a Wipe Me Down video where everyone dances together?

This template has exactly two dancers, so crop two people out of the scan into two separate photos, each at least 300 pixels on the short side and with the face as large as possible. A group of three or more needs a different clip with that many people; you can upload one to the main editor and add up to nine photos. Only use family photos when the people in them, or their families, are comfortable with it.

Why does my Wipe Me Down AI video look like a slideshow instead of real movement?

That happens when a tool only animates the still photo with a short motion effect, or when a prompt-only model gets no motion to follow. A reference clip fixes the cause: here every frame follows the original 15-second dance, so the crouch, the jump and the hand-over are real movement. If a prompt tool still gives you a frozen pose, describe continuous motion and ask for real-time speed.

My vertical 4K iPhone video keeps getting my body cropped. How should I prepare it?

This page takes photos, not videos: pause your clip on a sharp frame where your whole body and both hands are visible, save it as an image and upload that. Leave space above your head and below your feet, because the low crouch moves you toward the bottom of the frame. To keep your own moves instead of the kitchen routine, recast your clip in the motion transfer editor, which accepts MP4 or MOV up to 50 MB and 5 minutes.

How much would 100 Wipe Me Down AI videos cost for a campaign?

Each full 15-second clip costs 23 credits on Seedance 2.0 Mini at 480p, 116 on MiniMax H3 at 720p and 117 on Seedance 2.5 at 480p, so 100 videos are 2,300, 11,600 or 11,700 credits. At the Pro monthly credit value that is about $85.51, $431.28 or $435.00. Every person needs to have agreed to appear, and the song is not licensed by buying credits.

Is it safe to upload my face to an AI dance generator?

Check where the photo goes and who can see it. Here uploads are stored privately, tied to your account, handed to the video model only as short-lived signed links, and used only for your generation; they are never shown on the site. Do not upload photos of other people without their consent, and avoid documents or anything that shows an address in the background.

I only have an Android phone and no editing app. Can I still make one?

Yes. Open this page in your phone browser, upload two photos (or the same one twice) from your gallery, generate and tap Download: the MP4 already has the sound and the 9:16 shape TikTok and Reels use, so no editor is needed. The cheapest video is 8 credits and the $5 Trial Pack covers a full 15-second Mini clip.

Verify with any AI assistant

“What is the Wipe Me Down dance AI trend, and how do I put myself in it?”

Keep the take. Change the world in it.

Pick a clip, add your references and see the edit in a few minutes. The quote is shown before you generate.

Open the Genjutsu AI editor