Duo · 21sTwo photos take both guests
Both photos are replaced: you take the disc and your second photo walks in with the cup. The cafe, the caption and the sound stay as they are.
First Time in a Fancy Cafe AI Trend
Your photos take the wall disc until a guest sets her cup on it and shows it is a table.
Example · Duo · 21s
PNG / JPG / WebP · up to 10 MB · at least 512px on the short side
0/2 photos
Photo 1 is the man on the disc, Photo 2 is the woman who walks in. One person or one pet per photo, face clear, and Photo 1 head to toe if you can.
Video settings
Upload yourself, your own pet, or people who agreed to be in it; celebrity photos and film stills are not allowed.
Why the fancy cafe trend is a table, not a seat
First Time in a Fancy Cafe AI is a one-take POV clip set in a bare concrete cafe, where the small round disc on a curved metal post reads as a stool: the guest sits on it with his coffee until a woman walks in, sits on the bench and puts her cup on the same disc — which turns out to be a table.
Two discs hang above a continuous concrete bench, and at a glance they look like seats, so nobody in the clip is being careless: he is using the room the way the room advertises itself. The clip is also called your first time in a fancy cafe or first time at a fancy coffee shop, and it travels because everyone has walked into a room and guessed the furniture wrong.
The replies split on it. The top comment under a repost with about 6.9 million plays asks “what if she’s the one seating incorrectly”, with about 96,800 likes, and “I would remain seated to assert dominance” collects about 32,800 below it.
The four days the cafe clip spread
The earliest version that can be dated is set in another cafe, with magazine shelves and an orange light strip: @stevenwommack posts it on TikTok, and it reads about 2.1 million plays on November 3.
The concrete-cafe take is reposted with both people swapped for photos: @leehoustonx posts it at 15:38 UTC, and by November 3 it reads about 6.9 million plays.
Kevin Winnik (@kevkevkiwi) posts the concrete-cafe take itself on TikTok with the Steve Lacy sound, and it reads about 638,000 plays a day later.
He also posts a walkthrough of the settings behind it: Higgsfield, Seedance 2.5, a 22-second vertical clip built from two selfies, which is why this cafe is generated rather than filmed.
Play counts and comment likes read on TikTok on November 3, 2026. Posts are named by account; searches that add ai, template or how to return nothing yet.
The uncut take: sip, entrance, cup on the disc, climb down
One locked camera and no cuts for 21 seconds. The clock below runs from the first frame of the template.
0:000:20
10:00–0:05The first sip on the wall disc
The guest on the left sits on the small disc mounted above the bench, feet planted on the bench, and drinks his coffee with his back straight. Nothing about it looks wrong yet.
20:05–0:07A second guest comes in from the right
She walks in with her own cup, crosses the concrete floor and stops in front of the matching disc.
30:07–0:09She takes the bench and puts the cup on the disc
She chooses the bench, not the disc, and sets her cup down on the disc. It holds the cup flat, the way a table does.
40:09–0:15He looks over, then back at the camera
For six seconds nothing happens except his head turning towards her cup and away again.
50:15–0:20He climbs down and sits like he planned it
He stands up on the bench, steps down to the floor, sits beside her and puts his cup on his own disc. Both stare straight ahead until the clip ends.
One photo or two: who the take replaces
Both versions run the same cafe from the same locked camera and end on the corrected seat; only the number of photos changes.
Duo · 21s
Solo · 21sOne photo sits on the disc
One photo replaces the man on the disc; the woman who walks in, the cafe, the caption and the sound stay as they are, from the first sip to the last stare.
Who belongs on the disc, and what to caption it
Two photos make it a two-hander: one person gets the furniture wrong and the other gets it right in front of them.
The partner who orders for you
One photo puts them on the disc; add a second and you are the one walking in with a cup you did not choose.
Caption idea: he ordered for me again
The friend who moved here last month
Two photos: the local and the newcomer, and only one of them knows which disc is a table.
Caption idea: she has been here three weeks
Your dad at a wedding
One photo is enough: put him on the disc and let the guest in the clip be right in front of him.
Caption idea: “I’ll just wait here.”
The dog who treats every ledge as a seat
A pet photo can take the disc, with the cup set down beside it and the climb swapped for a hop up onto the bench.
Caption idea: he has never sat on a floor
How to make the First Time in a Fancy Cafe AI video
The cafe, the camera, the caption and the timing come from the template, and the photos decide who is in it. The roles are not tied to gender — put anyone on the disc and anyone in the doorway, or swap the two photos to swap who gets it wrong.
Photo 1Man on the table
Photo 1 is the man on the table: he is on the disc for the first sip, takes the sideways looks and climbs down at the end. Solo asks for this photo only.
- Full body if you have one: the first frame already has you up on the disc.
Photo 2Woman walking in
Photo 2 is the woman walking in: she sits on the bench and puts her cup on the disc, which is what turns it into a table.
- Only the duo version uses it.
- Waist up is enough; the clip keeps her on the right.
Photos that survive a wide shot
Use
One person per photo, head to shoes
A clear face, turned to the camera
A pet, whole body, looking at the lens
Avoid
Two people in one frame
Seated or cropped phone snaps
Dark or blurry photos
From one photo to a posted clip
Add one photo, or two
One for the solo version, two for the duo version.
Click Generate
The button shows the exact price for the resolution you pick before anything runs.
Post it
Download the MP4 and post it; the caption and the track are already inside.
Fancy cafe prompt tutorial
A kling prompt that edits the one-take cafe clip in place, duo or solo, plus the lines that keep the discs, the bench and the caption. The prompt this page runs is already set.
Kling 4.0 video edit
Prompt length 1131
@Video1 is the source clip: a 9:16 single take, locked-off camera, in a bare concrete cafe with two small discs on curved posts above a continuous bench. @Image1 and @Image2 are the two guests. Replace the man perched on the left disc, white short-sleeve shirt, black trousers, feet on the bench, with the person in @Image1; replace the woman who walks in from the right in the dark suit with the person in @Image2. If @Image1 is a pet, it takes the disc as a whole animal with four legs and no human hands or clothing.
Keep the take exactly as it plays: the sip, the slow head turn towards the other disc, the stand up on the bench, the step down to the floor, the seat beside her, the cup set on his own disc, the straight stare at the end, and the timing between them. Keep the locked camera, the framing, the single take with no cuts, the lighting, the composition and the pacing. Keep the burned-in caption "POV: Your first time in a fancy cafe" and the clip's own track unchanged. Do not re-shoot, redesign the cafe, move the discs or the bench, add people, text or logos, or let a face swap or blend between the two bodies.What the prompt has to say
Say what must not change: the discs, the bench, the caption
The joke lives in that furniture, and a re-shoot guesses it wrong.
Describe the woman by her suit and bag
It lets a solo prompt leave her alone and a duo prompt keep the two faces apart.
Keep “one take, no cuts” in the prompt
A version with cuts loses the moment where the cup lands on the disc.
What breaks it
“Redesign the cafe”
The bare concrete room is the setup, not decoration.
“Remove the caption”
It is burned into the clip and is part of the trend’s name.
“Make it 30 seconds”
The take is 21 seconds; length is not a prompt field.
One line, two ways
Asks for a new cafe
Put me in a fancy cafe video sitting on a table.
Edits the take and keeps the joke
Edit @Video1 in place: replace the guest on the left disc with @Image1, and keep the woman, the bench, the caption and the track.
The joke is the geometry: a disc that looks like a stool turns out to be the table. A re-shoot guesses the furniture and loses it.
Two versions: what each one asks for and costs
Two versions: what each one asks for and costs
- Photos
- 1 in the solo version, 2 in the duo version
- Photo files
- PNG / JPG / WebP, up to 10 MB, at least 512 px on the short side
- Length
- 21 seconds
- Frame
- 9:16 vertical
- Resolution
- 720p or 1080p; 720p by default
- Sound
- The clip’s own track, kept in the result
- Caption
- “POV: Your first time in a fancy cafe” stays on screen in every version
- Credits
- Solo · 21s: 789 credits at 720p, 1183 credits at 1080p; Duo · 21s: 789 credits at 720p, 1183 credits at 1080p
- Visibility
- Public on free accounts, private on paid plans
- If a run fails
- The credits go back to your balance automatically
More trending AI video effects
Every template works the same way: add your photos, press Generate and get a shareable video without writing a prompt.
View all effectsWhat people ask about the fancy cafe clip
Is the fancy cafe AI video free?+
No, each run costs credits. The price depends on the version you pick and the resolution, and the Generate button shows it before anything starts. New accounts start with 10 credits.
Is the First Time in a Fancy Cafe video real?+
No. The take was generated with an AI video model — the creator posted a walkthrough of the settings — and in your version the people are replaced by your photos.
What song is in the video?+
The clip carries its own track, listed on the TikTok post as Steve Lacy’s “oh yeah?”. It stays in the result, so there is no song to add afterwards.
Can my dog or cat sit on the disc?+
Yes, one pet photo can fill Photo 1: the pet takes the wall disc with the cup beside it, then moves to the bench for the ending. Animals are the input most likely to look off in a wide shot, so start from a clear full-body photo.
Is there a CapCut template for this?+
Yes, CapCut has edit templates for this trend. Here you add one or two photos and the finished take comes back with its caption and its track.
Can I change the caption?+
No. “POV: Your first time in a fancy cafe” is part of the clip and stays on screen in every version.
Can I change the cafe or the camera?+
No. The concrete room, the two wall discs, the bench and the locked-off camera come from the template; your photos change who is in the frame.
Why does my face look different?+
The camera is wide, so a face is a small part of the frame, and it turns away during the climb down. A full-body photo taken head-on keeps the proportions closer; a close selfie gives the model very little to work with, and a second run can land differently.
Do I need Higgsfield or Kling for this?+
No. Those tools ask you to set the prompt or the reference video up yourself; here the template carries the clip, the caption and the sound, and the only setting is resolution.
Can I get it in 16:9 or a square?+
No. The template keeps the clip’s vertical 9:16 frame, and resolution is the only setting you choose.
Is this the same as “first time at a fancy restaurant”?+
No. That is an older family of filmed restaurant skits with a waiter and a menu; this one is a single AI take about a wall disc that turns out to be a table.
Make your fancy cafe AI video
Add one photo to sit on the disc, or two to bring the guest; the cafe, the caption and the track come with the clip.