Built-in dance reference
The movement reference supplied with every generation. Your portrait replaces the performer in the request.
How does the iPhone Kupi video template work? See the built-in dance reference, what your photo changes, and how this AI workflow differs from a timeline edit.
By iPhone Kupi Dance AI · Updated October 28, 2026
This is a fixed reference-to-video workflow, not a downloadable editing timeline. Every request sends the same dance video and performance prompt to Wan 3.0. You supply one photo of the person who should appear in the new video.
The original Win William phone dance explains the format; the built-in reference is a separate 15-second AI remake. The song behind the trend is another distinct source.
| Part | Where it comes from | Can you change it here? |
|---|---|---|
| Performer | Your uploaded photo | Yes, choose a different photo |
| Dance, room and phone actions | Built-in video reference | No |
| Performance instructions | Built-in prompt | No |
| Model, duration and resolution | Wan 3.0, 15 seconds, 720p | No |
| Output crop | Adaptive to the vertical reference | No manual ratio selector |
No. It does not give you a timeline project or let you replace a clip in an editor. Wan 3.0 generates a new video from the photo and fixed movement reference, so faces, hands, phone shape, timing, dialogue, music and lip sync can vary. The reference is guidance, not a frame-by-frame copy.
Yes. The homepage publishes the prompt sent with the reference and photo. You can read or copy it there, but the generator does not have a prompt-editing field.
The movement reference supplied with every generation. Your portrait replaces the performer in the request.
An example of the iPhone Kupi Dance format. Your result may differ in appearance, motion, props and sound.