How to Make a Your First Time in a Fancy Cafe AI Video

Two still photos can recast the viral cafe sketch: one guest sits where a table is, the other walks in and uses the furniture the usual way. The joke is silent. Nothing needs to be said.

You do not write a prompt or edit a timeline. Upload two photos in the generator, generate a 21-second vertical clip, then download the MP4 for TikTok, Reels, or Shorts.

What the meme is

“Your First Time in a Fancy Cafe” is a locked wide shot in a concrete cafe. Someone perches on a small round metal table as if it were a chair. A second guest walks in, sits on the bench, and sets a cup on that table. The first person looks over, climbs down, and sits correctly.

On November 1, 2026, TikToker @leehoustonx posted an AI version set to Steve Lacy’s “Oh Yeah?”. It passed 9.2 million views in two days, according to Know Your Meme. Remakes keep the same cafe, camera, and joke, and swap the two people. On November 2, @kevinwinnik passed 14.5 million likes on Instagram, and @boby.biggsjohnson passed 4.2 million views on TikTok.

More on the origin and the trend: The Meme.

To recast the scene:

• Photo 1 · Wrong seat — already there, sitting on the table
• Photo 2 · Arriving friend — walks in, sits on the bench, puts a cup down

What you need

• Two photos, one person each, of consenting adults
• A sign-in and enough credits for one run

Watching the reference is free. Generating is not. The generator quotes the credit cost from length and resolution before you run. Failed generations return the credits spent.

Output is a 21-second 9:16 MP4. You can generate with sound or silent. There is no prompt and no watermark on the file.

How to make the video

1. Upload two photos. Photo 1 is the wrong seat. Photo 2 is the arriving friend. Both slots are required. JPG, PNG, or WEBP.
2. Generate. Sign in, check the quoted credits, then start. You can leave this page. Follow progress in My Videos.
3. Download and share. When the row is ready in My Videos, preview the MP4, download it, and post it. You are responsible for permission to use the faces, and for music or source-footage rights on the platform you use.

Swap the two uploads if you want to trade roles.

Which photos work

• Photo 1 · Wrong seat — one person, face clear, full body if you can
• Photo 2 · Arriving friend — one person, face clear, outfit visible

What actually helps:

• One person per photo, facing the camera with eyes open
• Face clear and well lit — no sunglasses, no hands over the face
• A full-body photo is best, so the whole outfit is in the shot
• No mirror selfies and no phone in hand — it can show up in the video
• Two different looks, so the roles stay readable in a wide shot
• Adults who agreed only. Do not upload minors, public figures, or anyone who has not consented

A face cropped from a group shot is usually too small once it has to fill a full-body figure.

The photos supply face, hair, and clothes. The cafe, camera, timing, poses, and the joke stay with the reference. Personal items from the reference stay in the shot, restyled to the person — including the arriving friend’s bag and the put-down.

Cost, time, and download

Each run is quoted in the generator. Credit packs are one-time; leftover credits stay on the account. See pricing for packs. This page is not a per-video price list.

When the clip is ready, preview and download it from My Videos. Failed runs refund the credits spent.

Short FAQ

Can I do this with one photo?

No. The scene has two people. You need two photos.

Is this a CapCut template?

No. It is an AI video generator. You do not edit the scene by hand.

Can I use a photo of a friend or partner?

Yes, if they are an adult and they agreed. Only upload yourself or consenting adults.

Do I add music myself?

You can generate with sound or silent. If you add a song in TikTok, Reels, or Shorts, that is between you and the platform.

Why did a face drift?

Usually a small crop, a dark photo, sunglasses, a hand over the face, or a mirror selfie with a phone in frame. Use a sharp single-person photo facing the camera and run again.

Not affiliated with @leehoustonx, TikTok, Instagram, or Steve Lacy.

If you already have the two photos, start from the generator.