Upload one photo of yourself, pick a scene and press Generate video: you sit on the wall disc, a regular walks in and sits properly, and it slowly clicks in about 10 seconds.
Seedance 2 Mini: text, frame and media reference video generation
Google's latest model with exceptional quality output
With one image, Grok follows the image aspect ratio.
Seedance and MiniMax also bill the actual duration of input videos. Documents: maximum 50 pages.
When this option is enabled, the output video may be selected by AI Generate and published to the Explore.
A first-timer on the wall disc next to a regular who sits the right way, and a couple on two wrong seats.
The samples start from original portraits of adults, not customers or celebrities. Up on the disc was cut from a 10-second vertical clip made with Seedance 2.0 Fast on November 3, 2026, which opened with him already in the cafe. Two first-timers, two wrong seats comes from another Seedance 2.0 Fast clip from the same day, with the couple up on the discs as the first frame.
About 9 seconds of each clip were kept for this page, gently reframed, colour graded and speed-adjusted, with the ghibliia.com mark at the top left. Nothing was mirrored. The music is original. Concrete Cafe and Sunset Terrace are image only: his same photo moved into two places with image-to-image AI, the same thing you can do on Image. No credits were spent on these samples. Use this photo →
A curly-haired guy in a mustard overshirt sips while perched on a steel disc on a concrete wall; a regular in a grey suit walks in with her cup, sits on the bench below, he looks down at her, climbs off the disc and sits on the bench like nothing happened.
Remix of a Seedance 2.0 Fast clip Use this photo
A woman with a braid and a guy in a blue jacket sip up on two wall discs while a server in an apron watches from below; they look at each other, climb down and end up on the bench, laughing.
Remix of a Seedance 2.0 Fast clip Use this photo
Image only The same guy sitting perfectly calm on the steel wall disc, cup to his lips, with the bench empty below him.
Still made in Image; this scene has no clip. Use this photo
Image only The same guy perched on a concrete planter on a terrace with string lights and an orange sky, cup in hand, the bench next to him empty.
Still made in Image; this scene has no clip. Use this photoUpload your photo, pick Concrete Cafe or Sunset Terrace, press Generate video.
Waist up or head to toe, facing the camera, sharp and well lit. Your clothes carry into the clip, so everyday clothes look most natural. The cafe comes from the clip, so any background works.
Above the photo box, choose Concrete Cafe, the steel wall disc from the trend, or Sunset Terrace, where the wrong seat is a tall planter. Same joke either way: you up high, the regular down where the seat is.
Press Generate video. The model, length and credit cost show on the button first. When it finishes, download the MP4. For another place or more people, write your own script on Video.
A proud sip on the disc, the regular sits right, the look, the climb down.
Here's the Up on the disc sample, about 9 seconds long:
0–2 s: he sips on the steel wall disc, completely at ease. A regular in a grey suit walks in with her cup and crosses in front of the camera.
2–6 s: she sits down on the bench below with her coffee. He stays up there, cup at his mouth, and glances down at her.
6–9 s: he eases off the disc and sits on the bench with his cup, staring ahead as if he'd been there all along.
Kevin Winnik and @leehoustonx both posted it on November 1, 2026; sources disagree on who was first.
Kevin Winnik (@kevkevkiwi) posted the clip with the song "Oh Yeah?" and got 588K views the same day. @leehoustonx's version passed 4 million views and 499.9K likes within a day, and #yourfirsttimeinafancycafe was created that day.
The first upload is disputed: some sources start with Kevin Winnik, others credit @leehoustonx with the filmed original and call the rest remakes.
On November 2 CapCut posted a one-photo template and AI versions followed, such as @kkubi99's AI-labelled take (259.5K views). One source counted well over a hundred creators posting it within three days.
Put the POV line on top, add the sound in TikTok, tag the friend who always orders the same.
The POV line goes on top: "POV: your first time in a fancy cafe". Then tag whoever it's about: the friend who orders a plain coffee, a dad loyal to the same diner cup, a first date, or yourself. Switch on the AI-generated label when you upload.
The clip comes with original music. In TikTok you can add the trend sound on top; for a beat of your own, make one on Music. More ideas in all video effects.
A TikTok POV where someone sips coffee perched on a small steel disc on the wall, treating it as a stool, until a regular sits on the bench and uses her disc as a table.
Kevin Winnik (@kevkevkiwi) and @leehoustonx both posted it on November 1, 2026, and the sources don't agree on the order.
One adult, waist up or full body, looking at the camera in good light.
Yes, CapCut released a one-photo template on November 2. Here you only need one photo in your browser.
The friend who always orders plain coffee, your dad, or yourself, as long as they're in on the joke.
