Dime Dime Baby AI Prompt: What to Write, and What a Prompt Alone Cannot Do
Updated 2026-10-22
Searching for a Dime Dime Baby AI prompt usually means you have a video model open and want to make the BbY WOW dance yourself. Here is a prompt that works, the settings around it, and the parts that a prompt cannot fix on its own.
First: you need a reference video, not just a prompt
Text alone will not get you the trend. "A man dances at a party" gives you a dance, not the dance. The Dime Dime Baby look comes from a specific clip — the strut down a quinceañera aisle, the snack bite, the deadpan close-up — so you need a model that accepts a photo plus a motion reference video (Seedance 2.x and similar "reference to video" models), and the clip itself, trimmed and cropped.
A prompt that works
Use your photo as image 1 and the trimmed clip as video 1:
Image 1 is the subject. Keep their face, hair, colours and body shape exactly as in image 1. Video 1 is the motion reference.
The subject from image 1 takes the place of the dancer in video 1 and recreates his whole performance: the same strut toward the camera, the same moves, timing, camera angle and framing. Everyone else in video 1 stays as they are.
0–2 seconds: a party hall with purple and pink lights. The subject walks toward the camera between the tables, doing a loose, bouncy shoulder-shuffle step, arms swinging. 2–4 seconds: still dancing forward, the subject takes a bite of a snack, chewing with total confidence. 4–7 seconds: medium close-up as the subject reaches the camera, chewing with a smug deadpan face, nodding to the beat.
Handheld phone footage with slight shake. The subject from image 1 is recognisable in every frame; only one of them, and no trace of the original dancer.
Settings that matter
| Setting | Use | Why |
|---|---|---|
| Aspect ratio | 9:16 | It is going to TikTok or Reels |
| Duration | 7 seconds | The whole dance; shorter cuts off the stare |
| Resolution | 720p | 480p loses the face on the close-up |
| Reference clip | Trimmed to the dance, letterbox cropped | Intros and black bars confuse the motion |
What usually goes wrong
- The original dancer leaks through. Say "no trace of the original dancer" and say it last.
- The face drifts by the close-up. Name the beats with timestamps; vague prompts let the model improvise the ending.
- Two of you appear. Add "only one of them".
- It fails the safety check. Real-person photos sometimes get blocked; a clearer, front-facing photo helps.
Expect a few paid attempts before one comes out right, plus the time to find, trim and crop the clip.
Or skip the prompt
The Dime Dime Baby AI page runs a tested version of this setup for you: the clip is already trimmed, the prompt is tuned, and the settings are fixed. You upload one photo, pay once with no sign-up, and get the video in about five minutes — including a Halloween edition until early November.