THE CAFE TABLE TREND

First time in a fancy cafe? Make your AI video.

Make it your moment.

Turn one or two photos into the fancy cafe AI video. Cast yourself, a friend or your pet in the tiny-table scene, choose your settings and generate your own clip. Your first video is free after sign-in.

5–15 seconds · 768p / 1K / 2K · First video free

MADE HERE · AI

THE PLOT TWIST

Generated here from two AI portraits and a pose reference

MiniMax-H315s · 768p · 9:16

Your cafe moment starts here

Choose a role, upload your photo and make the cafe moment. Your prompt is ready.

01Cast the scene

On the table

One clear face, good light, no heavy filters. Pets welcome.

02Choose your video

Length

Quality

Aspect ratio

Prompt & advanced settings

The scene prompt is filled in for your chosen role. You can generate immediately or customize it here.

Describe your scene

Image 1 identifies the first character, sitting awkwardly sideways on the table. Preserve face, hair, clothing or fur markings. The arriving guest is a different adult. A photorealistic, deadpan first-time-in-a-fancy-cafe comedy. Along a bare concrete wall is a continuous concrete bench about 45 cm high: this is the real seat. Two tiny circular stainless-steel coffee tables about 30 cm wide and 75 cm high project in front of it, supported by J-shaped metal tubes fixed into the bench's vertical front, with no floor legs. Keep this geometry fixed. The first character's BUTTOCKS rest directly on the left metal coffee tabletop, while BOTH SHOES rest on the LOWER CONCRETE BENCH behind it. Their hips are supported by the metal disk, and their feet by a different lower concrete surface. Knees are drawn up toward the chest in a cramped, fully clothed, toilet-seat-like crouched sitting posture. Torso and knees face sideways toward the wall/pillar on screen right, with back and rear shoulder toward the camera; the head may turn slightly to show a side profile. Hold the coffee near the raised knee. Clearly show these two separate support surfaces. Feet must not stand on the tabletop or dangle in front of it. Begin with the character seriously thinking this is the normal way to sit and taking one small coffee sip. Another guest enters from the right across the foreground, turns and sits normally on the concrete bench behind the right coffee table, feet toward the floor, coffee on the tabletop and hands resting around it. Keeping the initial raised-knee pose, the first character looks sideways at the guest's proper seating, then at the coffee on the tabletop, then down at the disk beneath their own hips. Hold a mortified pause: pressed lips, stopped sipping and averted eyes. The guest gives one brief, restrained sideways look and returns to facing forward. The first character then leans forward with the cup held steady, transfers weight to both feet on the concrete bench, lifts their hips off the metal disk and rises on the bench; carefully steps from the bench down to the floor, turns to the normal direction, and sits on the bench behind the left table, finally setting the cup on it. End with both seated normally, tables holding coffee, first character sitting rigidly straight and pretending nothing happened. Fit the stages to the selected duration, preserving the awkward realization. Fixed medium-wide camera from slightly behind and to the side, both heads, shoes, tabletops, bent tubes, bench and some floor visible. Soft daylight, consistent faces and anatomy. No extra seats, no bathroom, no undressing, no grins or exaggerated laughter, no duplicated characters. Original quiet playful piano jazz, soft bass, brushed drums, subtle footsteps, cup sounds and cafe ambience; briefly pause the music at realization. No dialogue, vocals, subtitles or on-screen text. For a pet preserve natural animal anatomy.

Choose your input

Use public HTTPS links for video and audio references. Keep video and audio references within 15 seconds in total per type. Up to 5 references; no first/last-frame mixing. Media support still needs a manual generation check.

5–15 seconds · Standard channel · One video per generation. Portrait references use your selected framing.

1 video credit per clip · First video free

Submitting uploads your selected media and sends it to Metaso for generation.

My videos

9:16 verticalYou, a friend, or your petFirst video free

THE SAME CAFE. A DIFFERENT MAIN CHARACTER.

A little awkward. A lot of you.

Examples

MADE HERE · AI

YOUR PHOTO15s · 2K · 9:16 · ♪

A tiny table. A very awkward realization.

This site's demonstration

Generated November 5, 2026 · MiniMax-H3 · 15 seconds · native 1440 × 2560 · 9:16 · audio present. Two AI adult portraits and the linked pose reference were used. This demonstrates this one task, not every input or setting.

View pose reference Make a scene like this

REFERENCE · AI

YOUR PHOTO15s · 480p · 9:16

One photo. Main character energy.

Make a scene like this

REFERENCE · AI

YOUR PHOTO15s · 480p · 9:16

Bring your partner in embarrassment.

Make a scene like this

REFERENCE · AI

YOUR PHOTO15s · 480p · 9:16

Your pet has expensive taste.

Make a scene like this

The first clip was generated here from the two AI adult portraits and pose reference shown. Other clips are labelled references. This example does not establish pet results or a general success rate.

ONE PHOTO. ONE VERY AWKWARD COFFEE.

The table is tiny. The joke is universal.

You walk into a cafe that is a little too cool for you. That little round thing must be a stool, right? Then somebody sits down properly, and it all clicks. This is your invitation to play the person who gets it spectacularly wrong.

Get to know the trend

NO CAMERA CREW REQUIRED

From camera roll to cafe chaos.

Read the guide

Choose who gets the wrong seat

Pick On the table for the lead role, Walking in for the guest, or Both of us to cast a friend too.

Give your photo a good chance

Use one subject per photo, facing the camera. For a pet, include the whole body. Avoid cropped faces, sunglasses and dark rooms.

Set the scene before you commit

Your scene prompt is ready. Choose the length and resolution, then generate. Follow the saved task and download the finished video.

A CLEAR PRICE BEFORE THE PUNCHLINE

One-time credit packs

Start with one free video. Buy 3, 10 or 30 more with a one-time payment. No recurring subscription.

Pricing

Starter

^$4.90

one-time payment

3videos

5–15s · Up to 2K

The sweet spot

Creator

^$9.90

one-time payment

10videos

5–15s · Up to 2K

Studio

^$19.90

one-time payment

30videos

5–15s · Up to 2K

All packs are one-time purchases. Choose 5, 10, or 15 seconds and 768p, 1K (1080p), or 2K in the studio. Each generation uses one video credit.

FAQ

Before you take a seat

5–15 seconds · 768p / 1K / 2K · First video free

Get help

What is the fancy cafe video trend?+

It is a short visual joke: someone mistakes a tiny cafe table for a seat, watches another guest sit correctly, then quietly fixes their mistake. The fun is in that moment of realization.

Can I use one photo or do I need two?+

One photo is enough to select the person on the table or the guest walking in. Choose Both of us when you want to provide photos for both roles.

Can I put my dog or cat in the scene?+

Yes, the studio accepts a pet photo for the On the table role. A clear full-body image is the most useful starting point. Reference examples show the idea; generated results can vary.

Can I generate a video here right now?+

Each account receives one free video. One-time packs (USD / video credits): $4.90 / 3 · $9.90 / 10 · $19.90 / 30. Choose 768p, 1K (1080p), or 2K in the studio.

Are the example videos results from this site?+

Examples are labelled by origin. Clips marked Made here were generated through this site and are shown beside their source photos. Reference clips illustrate the idea. Generated results can vary.

What happens to my photos?+

Photos are previewed locally until you submit. Submission uploads them to site storage and shares their public HTTPS URLs with Metaso for generation. Task history and completed results are stored by this site.

Read how media is processed

Does my video include the original trend song?+

The scene prompt requests original background music, and the current Made here example has an audio track. Audio can vary between generations. A reference clip's song is not a promise that your result includes that recording, and this site does not grant rights to the original trend song.

My task is still processing. Should I submit again?+

Open the saved task in My videos and use Check / resume. A browser timeout does not prove that the generation failed. Keep the same task while its submission or result is being checked so you do not start another generation unnecessarily.

Read the troubleshooting guide

Is a returned video credit the same as a cash refund?+

No. A confirmed failed or cancelled task returns its reserved video credit once. Payment disputes, duplicate charges and unused-pack questions require a support request with the order number. A pending request is not a completed cash refund.

Read the credit and refund rules

Will the output have AI identifiers or commercial-use rights?+

Outputs may contain AI identifiers; inspect the complete result before sharing. Use photos with permission and follow the destination platform's disclosure requirements. Neither a reference asset nor a generated soundtrack gives you rights to someone else's likeness, footage or song. No blanket third-party commercial-use licence is promised.

A LITTLE READING WITH YOUR COFFEE