How to Make a Fancy Cafe AI Video with People
A tiny table, a coffee and a moment of realization: make your own version of the fancy cafe POV joke from one or two photos.
What happens in the trend?
In a minimalist cafe, one customer mistakes a small wall-mounted table for a stool. A second customer sets a coffee on an identical table, revealing the mistake. Our version asks the model to retell the scene with people in a single 15-second shot.
Make your video
Choose a scene
Open the Fancy Cafe maker and choose one of the two person scenes.
Add one or two photos
Use a clear, well-lit photo with one subject per image, with permission. Photo 1 sits on the table; optional photo 2 is the arriving customer. Without it, the model invents a companion.
Generate and review
Choose a quality, check the displayed credit cost, and generate with Seedance 2.0 Fast. Preview the result before downloading.

Créez la vidéo Fancy Cafe avec 2 photos
Vidéo de 15 secondes au format 9:16, avec son propre audio, à partir de 100 crédits. Aucun prompt requis.
Ouvrir le générateur Fancy CafeCreate your own version
Describe a fixed vertical shot of a concrete cafe with a bench and two matching wall-mounted round tables. Ask for the first subject to perch on one with coffee, the second to enter and place a cup on the other, and the first to look down, climb onto the bench and act normal. Specify the subjects' faces and keep them distinct. Both reference videos are silent, so the scenes request no audio.
The timing is creative guidance, not a frame-accurate guarantee. A full-body photo helps the model interpret the pose and outfit.
Improve the result
- Use a sharper photo if the subject's face drift.
- Show the full person if body shape or posture is inconsistent.
- Try another generation if the second person or the table action is unclear.
- Add a caption or a licensed platform sound after downloading if desired.
FAQ
Can I add two subjects?
Yes. Upload a separate photo for each person. The second is optional.
Is the trend song included?
Both person references are silent, so the scenes request no audio.
Is the action guaranteed?
No. The matching reference video guides the action alongside your photos and prompt; timing, identity and action can vary.


