The “Oouuh” duet, made in your phone browser
What You Want Me To Do AI Video on Your Phone
The “What You Want Me To Do” AI video is the Fredo Bang “Oouuh” ukulele duet: a 2018 kitchen clip in which Einer Bankz strums on the left and Fredo Bang sings on the right. The phrase is a line from the hook, not the song’s title. On this page you make it in Safari or Chrome on your phone: pick two photos from your camera roll, choose a length and download a vertical MP4 with the original sound. No app to install and no CapCut template to track down.
- Two photos, or one used twice
- Original sound
- iPhone, Android, iPad
- From 13 credits
Default: 23 s · Seedance 2.5 · 480p · 110 credits. A video that fails returns its credits.
Make the What You Want Me To Do duet
Photo 1 · Ukulele (left)
One person or pet per photo
Photo 2 · Singer (right)
One person or pet per photo
The same photo in both slots puts two copies of one person side by side. A pet can take either part.
JPG, PNG or WebP · up to 30 MB each · both sides 300–6000 px. Upload only people who agreed, or your own pets.
23 s · 480p · 9:16 · MP4
The original take’s ukulele and vocal are laid under your video.
110 credits per video
Without a purchase, videos carry a watermark. Any credit-pack purchase unlocks the clean original.
One golden retriever playing both parts
The same dog photo in both slots: one copy strums with its paws, the other sings and throws its head back on both hooks. The full 23 seconds with the original sound.
What the “What You Want Me To Do” AI video is
A trend named after a lyric: people search the line they remember, and the clip behind it is Fredo Bang’s “Oouuh”.
The scene is one phone shot from 2018. Ukulele player Einer Bankz stands on the left of a small kitchen and strums, while rapper Fredo Bang, on the right, sings the hook of “Oouuh” to the camera and tips his head back on the long notes. “What you want me to do” is one of the lines he sings, which is why so many people look the trend up by that phrase rather than by the title.
In the AI version two photos take over both parts. Most people who see it are scrolling on a phone, so this page covers the phone route end to end: which browser to open, how to get two usable photos out of a camera roll, which length to pick, what it costs and how to get a sharp vertical file back onto TikTok, Reels or Shorts.
Watch Fredo Bang’s official “Oouuh” videoNo CapCut template needed: three ways to make it on a phone
Tutorials under the trend usually send you hunting for a template. Here is what each route involves.
| Route | What you do | Both parts and the moves | Sound | The catch |
|---|---|---|---|---|
| A CapCut or PixVerse template | Find a creator’s template through TikTok tags such as #pioneertemplate, open it in the app, add your photos | Depends on the template | Depends on the template; often added in TikTok afterwards | Needs the app installed; template links get removed or don’t open in every region |
| A general photo-to-video app | Write a prompt describing the duet and animate one photo | Usually one person per video; the kitchen and the moves are improvised | Generated audio, not the take | It rarely looks like the trend without a lot of retries |
| This page in Safari or Chrome | Pick two photos, a length and a room, then generate | Left strums, right sings, following the 2018 take’s moves and timing | The original take, already in the MP4 | Costs credits; a video takes a few minutes |
Making it on an iPhone, Android phone, iPad or laptop
Everything runs in the browser, so the steps barely change between devices. Open the page in Safari or Chrome itself, not inside the TikTok or Instagram app, where sign-in and downloads can fail.
| Device | Picking the photos | Where the MP4 ends up |
|---|---|---|
| iPhone · Safari | Tap a photo slot, then Photo Library. Use a still photo, not a Live Photo’s motion. If a photo is refused as the wrong type, export it as a JPEG first; renaming a HEIC file does not convert it | Download MP4 opens Safari’s downloads; find it in Files → Downloads, open it and choose Share → Save Video to put it in Photos |
| Android · Chrome (Samsung, Pixel and others) | Tap a slot and choose Photos or Files. Check you’re picking the original image, not a small preview saved from a chat | The download notification or Files → Downloads; share it from there if it doesn’t show in the gallery |
| iPad · Safari | The same as iPhone; the larger screen makes the controls and the price table easier to read | Files → Downloads, then Share → Save Video |
| Laptop · any browser | Drag a JPG, PNG or WebP into each slot | Your Downloads folder; AirDrop or message it to your phone to post |
You can close the page while the video is made. Sign in again on any device and it shows up under your recent videos and in My videos.
Getting two good photos out of a camera roll
The first frame is a still made from your photos, and the video copies every face from it. A clear face matters more than a perfect outfit.
| What you have | What to expect | What to do on the phone |
|---|---|---|
| A normal photo of one person facing the camera | The closest likeness | Use it as it is; a photo that shows the clothes keeps the outfit |
| One photo with both people in it | Can’t be used as one upload | Duplicate it in Photos or Gallery, crop each copy to one person, and upload one per slot |
| A front-camera selfie at arm’s length | Good face, guessed body | Fine for the singer; for the strummer a photo that shows the arms and hands helps |
| A screenshot from a video | Works if the frame is sharp | Pause on a still moment, screenshot, then crop away the app’s buttons and captions |
| A low-resolution picture saved from WhatsApp | Soft faces can drift | Ask the sender for the original or use the largest copy; each side needs at least 300 px |
| Sunglasses, a cap or a face in shadow | Eyes and hair are guessed | Pick another photo, above all for the singer, whose face fills the hooks |
Accepted: JPG, PNG or WebP up to 30 MB, each side 300–6000 px. Photo 1 plays the ukulele on the left with its mouth closed; Photo 2 sings. The same photo in both slots puts two copies of one person side by side, and a pet can take either part.
10, 15 or 23 seconds, and which room
Every length starts at the beginning of the take. A longer video holds more of the hook; it is never slowed down.
| Length | Where it stops | Good for |
|---|---|---|
| 10 s | Before the big head-back “ooh”: the chin-hand sway, a grin and a raised fist | A quick story or a reply video |
| 15 s | Just as the first long “ooh” begins | A short TikTok or Reel that loops |
| 23 s | After the hook twice, hands clasped for the last note (Seedance 2.5 only) | The whole bit, as people know it from the trend |
The default room is an empty kitchen like the original. To put the duet in your own place, stand where you’d film, hold the phone upright and take a photo of the empty room; upload it under “Where do they perform?”. Anything in it that is not the room is left out.
How to make it on your phone in three steps
1. Open the page and pick two photos
In Safari or Chrome, tap Photo 1 for the ukulele player and Photo 2 for the singer. Use one photo twice for a solo, or crop two people out of one picture first.
2. Choose the length, quality and room
23 s on Seedance 2.5 holds the whole hook; 10 or 15 s works on every model. Keep the kitchen or upload a photo of your room. The credit price shows before you start.
3. Generate and save the MP4
A still of the pair comes first, then the duet with the original sound, usually within a few minutes. Tap Download MP4 and save it to Photos or Gallery.
What one video costs
Credits come from one-time packs that never expire; there is no subscription. The price depends on the model, length and resolution and is shown before you generate.
| Model | Length | Resolution | Credits | Value at Studio rate |
|---|---|---|---|---|
| Seedance 2.5 | 10 s | 480p | 50 | $3.29 |
| Seedance 2.5 | 10 s | 720p | 108 | $7.11 |
| Seedance 2.5 | 10 s | 1080p | 259 | $17.05 |
| Seedance 2.5 | 15 s | 480p | 73 | $4.81 |
| Seedance 2.5 | 15 s | 720p | 159 | $10.47 |
| Seedance 2.5 | 15 s | 1080p | 386 | $25.41 |
| Seedance 2.5 | 23 s | 480p | 110 | $7.24 |
| Seedance 2.5 | 23 s | 720p | 242 | $15.93 |
| Seedance 2.5 | 23 s | 1080p | 590 | $38.84 |
| Seedance 2.0 Mini | 10 s | 480p | 13 | $0.86 |
| Seedance 2.0 Mini | 10 s | 720p | 22 | $1.45 |
| Seedance 2.0 Mini | 15 s | 480p | 17 | $1.12 |
| Seedance 2.0 Mini | 15 s | 720p | 31 | $2.04 |
| Seedance 2.0 Fast | 10 s | 480p | 22 | $1.45 |
| Seedance 2.0 Fast | 10 s | 720p | 43 | $2.83 |
| Seedance 2.0 Fast | 15 s | 480p | 31 | $2.04 |
| Seedance 2.0 Fast | 15 s | 720p | 63 | $4.15 |
| Seedance 2.0 Pro | 10 s | 480p | 35 | $2.30 |
| Seedance 2.0 Pro | 10 s | 720p | 69 | $4.54 |
| Seedance 2.0 Pro | 10 s | 1080p | 168 | $11.06 |
| Seedance 2.0 Pro | 15 s | 480p | 50 | $3.29 |
| Seedance 2.0 Pro | 15 s | 720p | 102 | $6.71 |
| Seedance 2.0 Pro | 15 s | 1080p | 250 | $16.46 |
| MiniMax H3 | 10 s | 720p | 48 | $3.16 |
| MiniMax H3 | 15 s | 720p | 69 | $4.54 |
Values use the cheapest sold credit: $79 ÷ 1,200 credits. They are a share of a prepaid pack, not a cash price for one video. Smaller packs cost more per credit. The price includes the still of the pair.
| Question | Current answer |
|---|---|
| Free credits | Eligible new accounts get 1 credits valid for 3 days. The cheapest duet costs 13 credits. That alone does not cover a video. |
| Smallest pack for the default video | Popular: $12.99 for 150 credits. It covers 1 × the default 23-second duet, or 11 × the cheapest 10-second one. |
| Watermark | Videos made before your first purchase carry a moving site watermark. After any credit-pack purchase, new videos are clean and earlier ones can be unlocked without generating again. |
| If the video fails | If the duet cannot be finished, including a photo the model’s review refuses, every credit is returned automatically. |
| A finished video you want different | A finished video is not refunded for taste. Making it again costs the listed credits again, so check the photos and the length first. |
Posting it from your phone without losing quality
Apps re-compress whatever you upload, so start from the best file you have.
| Setting | What to do |
|---|---|
| Resolution | 480p looks fine in a chat; pick 720p or 1080p before generating if it’s going on TikTok, Reels or Shorts |
| The file | Upload the downloaded MP4 itself; a screen recording adds blur and the phone’s status bar |
| Frame | The video is already vertical 9:16, so no cropping is needed in the app |
| Sound | The take’s ukulele and vocal are in the MP4. The song belongs to its rights holders; if an app mutes it, add the official “Oouuh” sound from its library, and keep it out of ads |
| Watermark | Before any purchase the video carries a moving site watermark; buying any credit pack removes it from new videos and unlocks earlier ones |
| AI label | Turn on the app’s AI-generated content label when you post |
Whose photos to use, and what happens to them
Your camera roll is full of other people. Before you pick, check these.
Friends, partners and family
Ask before you post someone’s face, even as a joke. A private chat is one thing; a public account is another.
Coworkers
Get each person’s okay before a team or company post, and don’t take photos from a staff directory.
Children, celebrities and strangers
Leave them out. The parts are written for adults, and the video model’s review usually refuses celebrity photos and film stills.
Your uploads
Photos are stored privately in your account, used to make your video and sent to the video model provider for that job. We never publish them.
What You Want Me To Do AI video examples
Finished duets that follow the same 2018 take, with its sound. Your video gets the same moves and timing.
One golden retriever playing both parts
The same dog photo in both slots: one copy strums with its paws, the other sings and throws its head back on both hooks. The full 23 seconds with the original sound.
Two friends in a small kitchen
One friend keeps his eyes on the ukulele while the other sings to the camera with a hand at his chin. 14 seconds with the original sound; it stops before the second hook.
A poodle on ukulele, a grey cat singing
Two different pets: the poodle strums on the left, the cat reaches toward the lens and tips its head back on the held notes. The full 23 seconds with the original sound.
Phone questions about the What You Want Me To Do AI video
Can I make the What You Want Me To Do AI video on my iPhone without an app?
Yes. Open this page in Safari, pick a photo for the ukulele player and one for the singer, choose a length and generate. Nothing to install and no CapCut template. The MP4 lands in Files → Downloads; open it and use Share → Save Video to put it in Photos.
Does it work on Android phones like Samsung or Pixel?
Yes, in Chrome or Samsung Internet. The steps are the same; the finished MP4 goes to Downloads, and you can share it straight to TikTok, Instagram or YouTube from there.
Can I make it on an iPad?
Yes, in Safari, exactly as on an iPhone. It isn’t free: every video costs credits, from 13 for 10 seconds on Seedance 2.0 Mini at 480p.
I only have one photo. Can I still make it?
Yes. Use the same photo in both slots (there is a “Use this photo for both” button) and two copies of you play both parts, one strumming and one singing. A pet works the same way.
My selfie is a low-quality picture from WhatsApp. Will it work?
It can, if the face is clear and each side is at least 300 px, but soft photos make faces drift. Ask for the original file or use the largest copy you have, and avoid screenshots of screenshots.
Where does the video go after I download it?
On iPhone and iPad, Safari saves it to Files → Downloads; open it and choose Save Video to move it into Photos. On Android it goes to the Downloads folder or appears in the download notification. Your finished videos also stay under My videos on this site.
Is it safe to upload face photos from my phone?
Uploads are stored privately in your account, used only to make your video and sent to the video model provider for that job; they’re never published. Upload only yourself, your pets or adults who agreed.