Unforgettable Car AI

ONE PHOTO · BOTH PARTS · FREDO BANG "OOUUH" DUET

What You Want Me To Do AI Video from One Photo

The "What You Want Me To Do" AI video is the Fredo Bang "Oouuh" ukulele duet: a 2018 kitchen clip where Einer Bankz strums on the left and Fredo Bang sings the hook on the right. "What You Want Me To Do" is a line from that hook, not the song's title. Here one photo is enough: it plays both parts, so the video shows two of you side by side – one strumming, one singing – with the original sound.

1 photo plays both parts (or 2 photos, one per part) · people or pets · 10, 15 or 23 s · 9:16 MP4 with the original take's sound

From 26 credits ($0.72 in Creator Pack value). The full 23-second duet is 245 credits at 480p. One photo costs the same as two. First purchase $9.90; no subscription, no site watermark.

Compare credit packs
How many photos

Your photo becomes the ukulele player on the left and the singer on the right: two of you in one kitchen.

Your photo · plays the ukulele and sings

One person or one pet, face clear. A full-body photo gives the outfit down to the shoes.

300–6000 px wide and high, up to 30 MB. One person or one pet per photo, even light, no sunglasses. A photo of two people is not split for you: crop each person into their own photo first. Adults only – use your own photo or one from someone who agreed.

Model
Where the duet happens

The empty kitchen matches the original clip. Or upload an empty photo of your own room – the duet moves there. A vertical photo with floor space and nobody in it works best.

Resolution
Video length

10 s ends before the big "ooh", 15 s reaches the first head-back note, 23 s has the whole hook twice (Seedance 2.5 only).

480p · 9:16 · 23 s MP4 · the original take's sound

This video costs 245 credits. Charged when you start, refunded automatically if it fails.

Use photos you have permission to upload. Technical failures return the credits; making a different version is a new paid run.

What You Want Me To Do AI video examples

Four clips with the same kitchen framing, moves and sound as this template. The first one is the one-photo version: the same dog in both parts.

One golden retriever playing both parts

23s · one subject in both parts · original sound

What is the What You Want Me To Do AI video?

It is the name people search for the Fredo Bang "Oouuh" AI trend. The source is one continuous phone shot from 2018: Einer Bankz holds a ukulele on the viewer's left and strums, Fredo Bang stands on the right and sings the "Oouuh" hook, a hand at his chin, then his chest, then his head thrown back on the long "ooh". The AI version keeps that kitchen, the camera and every move, and puts new faces in both spots. People remember the line "what you want me to do" from the hook, so the trend took that name.

QuestionAnswer
Other namesFredo Bang Oouuh AI video, Oouuh AI trend, the ukulele kitchen duet
Original clipFredo Bang singing "Oouuh" with Einer Bankz on ukulele in a small kitchen, filmed in 2018
Fewest photosOne: the same photo plays the ukulele player and the singer
Who can starAdults and pets; a pet stands upright and strums with its paws
Prompt or template needed?No. The kitchen, the two parts, the moves and the timing are built in
ModelsSeedance 2.5 (default), Seedance 2.0 Mini, Fast and Pro, MiniMax H3
Length and frame10 or 15 s on every model, 23 s on Seedance 2.5; vertical 9:16 at 480p, 720p or 1080p (1080p on 2.0 Pro and 2.5)
SoundThe original take's ukulele and vocal, laid under the finished video
Price26 to 1330 credits per video; 245 for the full 23 s at 480p. Same price for one photo or two

Make the What You Want Me To Do AI video from one photo

The trend is a two-person bit, but most people have one good photo, not two. In one-photo mode your photo is sent for both parts: the model builds a still of two copies of you in the kitchen, then animates them with the original moves. Nothing is invented for the second spot.

  • Left copy: strums

    Holds the ukulele, eyes on the strings, mouth closed the whole time, with a couple of glances up at the camera.

  • Right copy: sings

    Hand at the chin, then the chest, a wide grin, a fist by the head, and the head thrown back on each long "ooh".

  • Both copies: your look

    Same face, hair, skin tone, build and clothes from your one photo. What the photo doesn't show – shoes, the lower half of an outfit – the model has to fill in.

What You Want Me To Do AI video examples

Four clips with the same kitchen framing, moves and sound as this template. The first one is the one-photo version: the same dog in both parts.

One golden retriever playing both parts

23s · one subject in both parts · original sound

Two friends in black T-shirts

14s · two subjects · original sound

A poodle strums, a grey cat sings

23s · two subjects · original sound

Plaid shirt on ukulele, tank top singing

9s · two subjects · original sound

Which single photo to use

With one photo, that photo carries both parts, so its framing matters twice. Here is what each common kind of photo gives.

The photo you haveWhat comes outTip
Full body, standingClosest match: both copies keep your whole outfit, down to the shoesBest choice. Arms away from the body help the strumming hands
Waist upFace and top stay yours; trousers and shoes are filled in to fitFine for most posts. Wear the top you want to see
Close-up selfieFace holds well; the body, outfit and sometimes a cap are inventedHold the phone further away, or use a mirror photo instead
Low-resolution chat photoWorks if it is at least 300 px on each side, but the face can softenAsk for the original file, or pick 720p so the face has more detail
Old printed photo or scanUsable; grain and colour cast carry over into both copiesScan flat in daylight, crop off the white border, one person per crop
Pet photoTwo of your pet: one strums with its paws, one sings with its own muzzleA side-on or standing pet photo shows the body; a face-only shot is guessed below the neck

Only have one photo of the two of you? Crop it into two

A couple photo, two friends at a party, or your parents in an old picture: the generator does not split a photo of two people, and it would read them as one subject. Turn it into two photos first, then use two-photo mode.

  1. 01

    Open the photo in your phone's editor and duplicate it, so you keep the original.

  2. 02

    Crop the first copy to the person who should strum: from the top of the head to as low as the photo goes, with nobody else's face in it.

  3. 03

    Crop the second copy to the person who should sing, the same way. Each crop must stay at least 300 px on its shortest side.

  4. 04

    Switch to "Two photos", put the strummer in Photo 1 and the singer in Photo 2. Use Swap to try the parts the other way round.

Faces cut off at the edge of the crop, or a shoulder of the other person left in, are the usual reasons a crop looks wrong. A small face in a wide group shot is better at 720p.

One photo, two photos or a cropped group photo

All three run the same template and cost the same. What changes is who appears and how much of each person the model can see.

What changesOne photoTwo photosOne group photo, cropped
Who appearsYou twiceTwo people or petsThe two people in the picture
What each part looks likeIdentical: same face and outfit on both sidesEach keeps their own face and clothesEach keeps their own face; outfits only as far as the crop shows
Who strums, who singsYour copy on the left strums, the one on the right singsPhoto 1 strums, Photo 2 sings; swap to tradeYou decide when you place the crops
SetupUpload onceUpload two photosCrop twice, then upload two
Main riskA close-up leaves both outfits to guessworkVery different lighting in the two photosSmall or cut-off faces from a wide shot
Price245 credits for 23 s at 480p245 credits for 23 s at 480p245 credits for 23 s at 480p

How other one-photo versions of this trend work

"One photo" means different things across the apps and templates people use for this trend. Checked 23 October 2026.

RouteWhat your one photo doesWhat you get
Single-image face-swap templateReplaces the ukulele player onlyYou strum next to the original singer
Two-portrait trend modeNot accepted: it asks for two separate portraitsNo video until you add a second photo
Group-photo or "original song" modeTakes one photo that already shows two peopleBoth people animated to the song, not always the kitchen moves
This page, one-photo modePlays both partsTwo of you in the kitchen duet with the original moves and sound

General photo-to-video apps can animate one photo of you singing, but they do not put a second performer with a ukulele next to you without a lot of prompting.

How to make it from one photo

  1. 01

    Upload one photo

    Keep "One photo" selected and add a photo of yourself or your pet, full body if you have one. It will play the ukulele and sing.

  2. 02

    Pick the place, model and length

    Keep the kitchen or upload your own room. Seedance 2.5 makes the full 23 seconds; every model makes 10 or 15. The price shows before you start.

  3. 03

    Generate and download

    A still of the two copies is made first, then the duet with the original sound. Download the MP4 and post it to TikTok, Reels or Shorts.

What You Want Me To Do AI video price: one photo or two

Every model, length and resolution. The price is per video and does not change with the number of photos or the background. Dollar values use the lowest-priced credit: the Creator Pack, $49.90 for 1,800 credits.

ModelLengthResolutionCreditsCreator Pack value
Seedance 2.510s480p110$3.05
Seedance 2.510s720p240$6.65
Seedance 2.510s1080p582$16.13
Seedance 2.515s480p162$4.49
Seedance 2.515s720p357$9.90
Seedance 2.515s1080p869$24.09
Seedance 2.523s480p245$6.79
Seedance 2.523s720p544$15.08
Seedance 2.523s1080p1330$36.87
Seedance 2.0 Mini10s480p26$0.72
Seedance 2.0 Mini10s720p49$1.36
Seedance 2.0 Mini15s480p36$1.00
Seedance 2.0 Mini15s720p70$1.94
Seedance 2.0 Fast10s480p49$1.36
Seedance 2.0 Fast10s720p98$2.72
Seedance 2.0 Fast15s480p70$1.94
Seedance 2.0 Fast15s720p145$4.02
Seedance 2.0 Pro10s480p78$2.16
Seedance 2.0 Pro10s720p161$4.46
Seedance 2.0 Pro10s1080p395$10.95
Seedance 2.0 Pro15s480p114$3.16
Seedance 2.0 Pro15s720p238$6.60
Seedance 2.0 Pro15s1080p589$16.33
MiniMax H310s720p109$3.02
MiniMax H315s720p160$4.44

The cheapest video is 26 credits (Seedance 2.0 Mini, 10 s, 480p). The default – Seedance 2.5, 23 s, 480p – is 245 credits, more than the 170-credit Single Reveal ($9.90); the Starter Pack ($19.90, 510 credits) covers two. MiniMax H3 renders at its 768P tier and delivers 720p.

ItemWhat applies
FreeBrowsing, the examples and the price table. No free video: the cheapest one needs 26 credits.
First purchase$9.90 buys 170 credits. Packs are one-time, never renew, and credits last 365 days.
WatermarkNone. The MP4 downloads clean, with the original take's sound, at no extra charge.
Failed videoA rejected photo or a failed generation returns the credits automatically.
Finished but not what you wantedNot refunded. Trying a different photo or model is a new paid run.
Compare credit packs

Where the What You Want Me To Do trend comes from

Einer Bankz is a ukulele player known for clips of rappers singing over his strumming. In 2018 he filmed Fredo Bang in a narrow kitchen singing the hook of "Oouuh"; Einer Bankz later re-posted the take to TikTok in 2022.

In early October 2026 the take turned into an AI trend: people dropped themselves, friends and pets into the two spots, mostly through CapCut or PixVerse templates shared under tutorial hashtags. Searchers call it by the line they remember from the hook.

Typing the phrase alone into a search engine mostly brings up general AI video makers or older songs with similar titles. The AI trend is this kitchen duet.

The 2018 kitchen take (Einer Bankz's TikTok post)Fredo Bang – "Oouuh" (official video)

This generator is independent and not affiliated with or endorsed by Fredo Bang or Einer Bankz.

One-photo questions

Can I make the What You Want Me To Do AI video with only one photo?

Yes. Choose "One photo" and upload it once. The same photo plays the ukulele player on the left and the singer on the right, so you appear twice in the kitchen. It works on a phone browser; no app or CapCut template is needed.

With one photo, which part do I play?

Both. The copy on the left strums with its mouth closed; the copy on the right does the singing, the grin and the head-back "ooh". If you'd rather only sing next to someone else, use two photos and put yours in Photo 2.

I only have one photo of the two of us together. Can I use it?

Not as one upload: a photo of two people is not split automatically. Duplicate it, crop one person into each copy (at least 300 px on the short side), then use two-photo mode with the strummer in Photo 1 and the singer in Photo 2. An old family photo of your parents works the same way.

Will a low-resolution selfie from WhatsApp work?

If it is at least 300 px on each side, yes, but chat apps compress photos and the face can come out soft. Ask for the original file if you can, choose 720p, and pick a photo where the face is large and evenly lit.

Is a one-photo video cheaper than a two-photo one?

No, the price is the same: it depends only on the model, length and resolution. From 26 credits (Seedance 2.0 Mini, 10 s, 480p) to 245 credits for the full 23 seconds on Seedance 2.5 at 480p.

Why did my clothes or hair change when I used a close-up selfie?

A close-up only shows the face and maybe a collar, so the model has to invent the rest of the outfit for both copies. Use a waist-up or full-body photo showing what you want to wear, and keep sunglasses and hats off unless you want them in the video.

Can I make a 15-second vertical video from one selfie on my phone?

Yes. Every video is 9:16, and 15 seconds is available on every model – it ends just as the singer throws the head back for the first long note. On Seedance 2.0 Mini at 480p a 15-second video is 36 credits; at 720p it is 70.

Can my pet play both parts from one photo?

Yes. Your pet appears twice, standing upright: one strums the ukulele with its paws, the other sings with its own muzzle. A photo that shows the whole animal gives the most natural result.