Pet edition · Fredo Bang "Oouuh" duet
What You Want Me To Do AI Video With Your Dog or Cat
The "What You Want Me To Do" AI video is the Fredo Bang "Oouuh" trend: the 2018 kitchen take where Einer Bankz strums a ukulele on the left and Fredo Bang sings on the right. The phrase is a line of the hook, not the song's title. Here your pet takes a side: upload a photo for the ukulele and one for the singer (the same photo twice gives one pet both parts), and the original take's moves and sound do the rest in your browser, with no CapCut template to hunt for.
Make my pet's version- Dogs, cats & owners
- 2 photos or 1 twice
- 10 / 15 / 23 s
- Original sound
23 s · 9:16 · 480p · original sound added after generation
Seedance 2.5 keeps fur, faces and paws best and is the only model for the full 23 s; Seedance 2.0 Mini is the cheapest way to test a pet photo. Credits are charged when the job starts and refunded automatically if it fails.
Your two photos and the room first become one opening frame, then the video follows the moves and timing of the original take (from a depth map, so none of the original people appear). The take's own sound is added after generation.
The pet version of the Fredo Bang "Oouuh" trend
Search "what you want me to do AI video" and you are looking for one clip: two performers in a narrow kitchen, a ukulele on the left, a singer on the right who tips the head back on a long "ooouuh". The pet version keeps all of that and swaps in your animals.
The original is a 2018 phone take of Fredo Bang singing "Oouuh" while Einer Bankz plays ukulele beside him, later posted on TikTok by Einer Bankz. "What you want me to do" is a line people remember from the hook, which is why the trend goes by that name. The original post on TikTok
When the trend turned into an AI template in October 2026, some of the versions people shared had no people at all: a golden retriever strumming next to itself, a poodle on ukulele with a grey cat singing. Pets suit it because the moves are simple, the camera never cuts and the hook is one big head-back moment.
This generator does not animate a still photo on its own. It follows a depth map of the original 23 s take, so your pets get the same sway, glances, hand-to-chest moments and the head-back hook at the same seconds, then the take's own sound is laid back under the video.
| Question | Answer |
|---|---|
| What it is | An AI remake of the 2018 Fredo Bang "Oouuh" kitchen take: a ukulele duet with a head-back hook. |
| Why "What You Want Me To Do" | It is a line of the hook, not the title. The song is "Oouuh". |
| Who does what | Photo 1 strums on the left and never sings; photo 2 sings on the right and does every move. |
| Can pets do it | Yes: dogs, cats and other animals stand upright, hold the ukulele with their own paws and sing with their own muzzle. |
| What stays from the original | The moves, the timing, the slightly handheld camera and the original sound. |
| Photos needed | Two, or the same photo twice for one pet in both parts. |
| Room | An empty kitchen like the original, or a photo of your own room. |
| Lengths | 10 s or 15 s on every model; the full 23 s (the hook twice) on Seedance 2.5. |
| Models | Seedance 2.5 (default), Seedance 2.0 Mini, Fast and 2.0, MiniMax H3. |
| Output | Vertical 9:16 MP4 at 480p, 720p or 1080p, with sound. |
| Cost | From 19 credits (10 s, Seedance 2.0 Mini, 480p); the full 23 s on Seedance 2.5 is 183 credits at 480p. |
| Watermark | On videos from free accounts; any paid order removes it. |
Pet and people versions of the same take
All four follow the same kitchen take: same layout, same sides, same moves and the original sound. Watch where the head goes back around 14 seconds; that is the moment pet versions are shared for.
One golden retriever, both parts
23 s · 9:16 · original sound
One dog in both parts: the left copy strums, the right copy sings and howls the hook, paws and all.
A poodle on ukulele, a grey cat singing
23 s · 9:16 · original sound
Two different animals, each with its own coat and face; the poodle wears a collar with a name tag.
Two friends in the original kitchen
14 s · 9:16 · original sound
The people version, for comparison: the strummer looks down at the strings, the singer gestures at the camera.
Plaid shirt on ukulele, tank top singing
9 s · 9:16 · original sound
A shorter 9 s cut of the opening, before the hook.
Your result follows the same take, so expect the same framing and timing; how close it looks to your pet depends mostly on the photo (see below).
Who strums, who sings: casting your pets
Photo 1 always takes the ukulele on the left and photo 2 always sings on the right. The singer gets nearly all the movement and the head-back hook, so give that side to the more expressive animal, or the one whose face you most want people to see.
| Cast | Photo 1 · ukulele | Photo 2 · singer | What to expect |
|---|---|---|---|
| Two pets | The calmer pet | The pet with the most expressive face | Each keeps its own fur and markings. Use photos taken from a similar distance so the two come out at believable sizes. |
| One pet, both parts | Your pet's photo | The same photo again | Two copies of the same animal, one strumming and one singing, like the golden retriever example. |
| Your pet plays, you sing | Your pet | You | You do the singing, gestures and head-back hook; your pet strums beside you on its hind legs. |
| You play, your pet sings | You | Your pet | The funniest version for most people: you strum seriously while your dog or cat howls the hook. |
| Dog and cat | Usually the dog | Usually the cat | A cat singer reads well because the head-back pose suits it; swap the photos if your dog is the louder one. |
Wrong way round after you uploaded? Use the swap button between the two photos before you generate.
Which pet photo works
The first frame is built from your photos, and the video copies that frame for every second that follows. A photo that shows the animal clearly is worth more than a higher resolution or a stronger model.
| Your photo | What happens |
|---|---|
| Whole body, standing or sitting up | Best case: the model sees the legs, paws, chest and tail it has to stand upright and strum with. |
| Lying down or curled up | Works, but the body and legs are invented; expect proportions to drift more than the face. |
| Face close-up only | The face usually holds; the body, size and coat pattern below the neck are guessed. |
| Fur close to the background colour | Edges blur and patches of coat can merge into the room. Pick a photo with some contrast. |
| Collar, tag, harness or a sweater | Small items like a collar and tag tend to stay; the model is told not to add clothes, so a bulky outfit may come out simplified. |
| Two pets in one photo | Crop it twice so each upload shows one animal; one photo with two animals in it can produce extra animals. |
| Motion blur, zoomed-in screenshot | The face is rebuilt from what is visible; markings and eye colour are the first things to change. |
| Black or dark fur in dim light | Details such as eyes and muzzle shape get lost; a daylight photo keeps them. |
What the model is told to do with an animal
- Stay a real animal standing upright on its hind legs
- Keep its own face, muzzle, fur and markings in every frame
- Hold and strum the ukulele with whole paws, never human hands or fingers
- Sing with its own muzzle, rounded open on the held notes
- No added clothes
- Exactly two performers, nobody else in the room
Only one pet and want it on its own, not in a duet? Burning Bridges AI takes one photo of a person or a pet.
When a pet duet goes wrong, and the fix
Most pet problems come from the photo or the casting, not from the model. Change the input first; switch models second.
| You see | Likely cause | Try this |
|---|---|---|
| Your dog strums with human fingers | The front paws are hidden in the photo, so the model has nothing to copy on the strings. | Use a photo where the front legs and paws are visible, and Seedance 2.5, which keeps paws best. |
| The pet's face drifts toward a person's | A small or soft face in the photo, often with a person next to it. | Crop to the pet alone so its face fills more of the picture; avoid photos where a person is touching it. |
| Two pets look like the same animal | Similar breeds or colours, or one blurry photo. | Use two sharp photos with the markings visible; put the one with the clearer face on the singer side. |
| The muzzle stays shut on the hook | A closed-mouth photo on the singer side and a short clip. | Pick 15 or 23 s so the hook is in the video, and a singer photo with the mouth slightly open. |
| The ukulele bends, floats or changes size | A very small animal for the instrument, or a lying-down photo. | Give the ukulele to the bigger animal, or to yourself, and the singing part to the small one. |
| The wrong pet is singing | Photo order: photo 1 always strums, photo 2 always sings. | Swap the two photos and generate again. |
| A tiny pet comes out much bigger than in life | Each performer takes the standing spot of one of the original pair; the photo sets the look, not the scale against the room. | Expected with this take. To show a real size difference, cast yourself in the other part. |
| Your room changes or gains furniture | A cluttered or angled room photo. | Use an empty, straight-on, upright photo of the room, or keep the kitchen. |
If a job fails, or the provider refuses a photo, the credits come back automatically. A video that finishes is charged even if you would have preferred another take, so test a new pet photo at 10 s on Seedance 2.0 Mini before spending on 23 s.
Which animals work
The template is built for two upright performers, one holding a ukulele. Animals that can plausibly stand and hold something give the most natural results.
Dogs
The classic pet version. Any size works; long-haired and curly coats keep their texture well when the photo is sharp.
Cats
Great singers: the head-back hook suits a cat. Expect a calm, upright cat rather than a playful one.
Rabbits, guinea pigs and hamsters
They become life-size performers. Put them on the singer side; a full-body photo helps the model place the paws.
Birds
A parrot has no paws to strum with, so the model has to improvise. Use a bird as the singer and test at 10 s first.
Stuffed toys and figures
Treated like an animal or a figure with arms. Plush toys with clear faces work best.
Horses, reptiles and fish
Possible, but far from the original's shape; results vary a lot. A short test costs little.
Length, the hook and which model to pick
The take never cuts, so a shorter video is simply the opening of it. What changes between lengths is whether your pet reaches the head-back hook.
10s
The sway, the grin and the first fist, no hook. Cheapest way to see whether a pet photo works.
15s
Ends just after the head-back hook starts at 14 s. Enough for a quick share.
23s
The whole hook (14–17.5 s) and the second round of the line, ending on the final held note. Seedance 2.5 only.
| Model | Lengths | Resolutions | For a pet duet |
|---|---|---|---|
| Seedance 2.5 | 10s · 15s · 23s | 480p · 720p · 1080p | Default. Keeps fur, markings and paws best; the only model for 23 s. Photos are checked by the provider first. |
| Seedance 2.0 Mini | 10s · 15s | 480p · 720p | Cheapest. Good for testing whether a photo reads as your pet. |
| Seedance 2.0 Fast | 10s · 15s | 480p · 720p | A middle step between Mini and the full models. |
| Seedance 2.0 | 10s · 15s | 480p · 720p · 1080p | Sharper detail than Mini and Fast, up to 1080p. |
| MiniMax H3 | 10s · 15s | 720p | A different look and motion; 720p only. |
Under the hood this is an AI video-to-video generator: the original take, as a depth map with no faces or room, supplies the motion and timing, while your photos and the opening frame supply every look. That is why the moves line up with the sound when it is laid back.
How to make the What You Want Me To Do AI video with your pet
1. Add the two photos
Photo 1 for the ukulele on the left, photo 2 for the singer on the right. Use the same photo twice for one pet in both parts. JPG, PNG or WebP, at least 300 px on the short side.
2. Keep the kitchen or use your room
The empty kitchen is selected by default. To use your own room, pick "Your room" and upload an empty, upright photo of it.
3. Choose length, model and resolution
Test a new pet photo at 10 s on Seedance 2.0 Mini, then make the full 23 s on Seedance 2.5. The credit cost updates as you choose.
4. Generate and download
Your photos and the room become the opening frame, the video follows the original take, and the original sound is added. Download the 9:16 MP4 when it is ready; it stays in your recent videos.
Credits for a pet duet
Every combination below is the exact quote the generator charges. A 10 s test on Seedance 2.0 Mini at 480p is 19 credits; the full 23 s on Seedance 2.5 is 183 credits at 480p, 405 at 720p and 992 at 1080p.
| Model · resolution | 10s | 15s | 23s |
|---|---|---|---|
| Seedance 2.5 · 480p | 82 | 121 | 183 |
| Seedance 2.5 · 720p | 179 | 266 | 405 |
| Seedance 2.5 · 1080p | 434 | 648 | 992 |
| Seedance 2.0 Mini · 480p | 19 | 27 | — |
| Seedance 2.0 Mini · 720p | 36 | 52 | — |
| Seedance 2.0 Fast · 480p | 36 | 52 | — |
| Seedance 2.0 Fast · 720p | 73 | 108 | — |
| Seedance 2.0 · 480p | 58 | 85 | — |
| Seedance 2.0 · 720p | 120 | 178 | — |
| Seedance 2.0 · 1080p | 295 | 440 | — |
| MiniMax H3 · 720p | 81 | 120 | — |
At the cheapest monthly-plan credit rate ($0.037), the 10 s test works out to about $0.71 and the full 23 s on Seedance 2.5 to about $6.80 at 480p or $36.88 at 1080p. That is a credit rate, not a per-video purchase.
The $5 trial pack (24 credits) covers one 10 s test on Seedance 2.0 Mini at 480p. The Starter plan (240 credits for $9 a month) covers the full 23 s on Seedance 2.5 at 480p.
Free account, watermark and refunds
- New accounts may receive 5 signup credits to try it.
- Videos made on a free account carry a moving raindancemaker.com watermark.
- Any paid order removes the watermark from new videos and unlocks the clean originals of the ones you have already made.
- Credits are charged when a job starts. If it fails, they are refunded automatically.
- The opening frame made from your photos is included in the quote; there is no separate image charge.
Compare plans on the pricing page
Resolution sharpens a good photo; it does not rescue a blurry one. Fix the photo before paying for 1080p.
Credits for your pet's duet
Every plan works for this generator and every other template on this site. Any paid order removes the watermark from new videos and unlocks the clean originals of earlier ones.
Starter
For trying the trend
240 credits / month
≈ 1 ten-second MiniMax H3 720p videos
- Watermark-free videos, including ones you already made
- Failed videos refunded
- Email support
Pro
For regular creators
780 credits / month
≈ 4 ten-second MiniMax H3 720p videos
- Watermark-free videos, including ones you already made
- Failed videos refunded
- All lengths up to 15 seconds
- Priority support
Enterprise
For agencies and pages
2,650 credits / month
≈ 14 ten-second MiniMax H3 720p videos
- Everything in Pro
- Dedicated support
- Custom integrations
Pet questions about the What You Want Me To Do AI video
Can I put my dog or cat in the What You Want Me To Do AI video?
Yes. Upload your pet as photo 1 to have it strum the ukulele, or as photo 2 to have it sing; the animal stands upright, keeps its own fur and face and uses its own paws. A 10 s test costs 19 credits and the full 23 s on Seedance 2.5 costs 183 credits at 480p. New accounts may receive 5 signup credits to try it.
Can one pet play both parts?
Yes. Upload the same photo for both slots and you get two copies of the same animal: one strumming on the left, one singing on the right.
Can I be in the video with my pet?
Yes. Put your own photo in one slot and your pet in the other. The singer side gets the gestures and the head-back hook, so most people give that part to the pet.
Why did my dog get human hands?
Usually because the front paws are hidden in the photo, so the model has to invent what holds the ukulele. Use a photo that shows the front legs and paws, keep Seedance 2.5, or put the dog on the singer side, which doesn't have to hold anything.
My pet doesn't look like my pet. Which photo works best?
A sharp daylight photo of one animal, whole body visible, face not too small, with fur that stands out from the background. Crop group photos so each upload shows one pet. Low-resolution or blurry photos lose markings and eye colour first.
Does it work with rabbits, birds or stuffed animals?
Rabbits, guinea pigs and plush toys usually work, best as the singer. Birds have no paws to strum with, so results vary; test at 10 s before making a longer version.
Does my pet need to be standing in the photo?
No. The model stands every animal upright on its hind legs. A standing or sitting-up photo just gives it more of the real body to copy, so proportions stay closer.
Can I post my pet's version on TikTok with the original sound?
The MP4 already carries the original take's sound. For a personal post that is how most versions are shared; for an ad or a brand account, replace the audio with music you have a licence for, because the song itself is copyrighted. Label the video as AI-made where the platform asks.
Give your pet the ukulele
Two photos, or one photo twice. Your pet in the kitchen duet with the original sound; start with a short, cheap test.
More one-photo and pet-friendly templates: Burning Bridges AI Video · All templates