What happens in the window dance AI video, second by second?
The AI builds the whole scene from your photo in one continuous shot, like a phone video filmed from across the street:
| Time | What you see |
|---|---|
| 0 to 2 s | Night. A dark brick apartment block with one warm, lit window. The camera flies slowly toward it until the window fills the frame. |
| 2 to 7 s | You dance alone behind the glass next to a small speaker on a wooden table: bounces, fist pumps, arm swings and a big grin, like nobody is watching. |
| 7 to 10 s | You notice the camera, slow down to a little shoulder groove and smile straight into the lens, caught. |
| 10 to 12 s | Still smiling, you grab the curtain and pull it across until the window is covered. The curtain stays softly lit. |
How it works: top-quality AI video models, including Seedance 2.5, make the clip straight from your photo, with no in-between picture, so your face stays yours. You get 12 seconds, vertical 9:16, with sound: an upbeat beat from the speaker, muffled through the glass, night sounds and the swish of the curtain. The music is new in every clip, not the trend's song, so you can add the trend's sound when you post. The exact price shows on the Generate button before you spend anything, and every run is a new take.
The Pet version plays the same night shot with your dog or cat up on the table next to the speaker. It stands up on its hind legs and dances to the beat, bobbing its head and waving its front paws (2 to 7 s), then spots the camera, freezes, drops back onto four paws and stares straight into the lens (7 to 10 s). The stare is the ending: no curtain. It runs 10 seconds, with the same night sounds and music.

