The celebrity kitchen AI video is a 15-second AI skit in which two famous-looking people fail to make one cup of coffee: one spoons it onto the counter, the other pours milk all over it. videosi makes your own version from one photo, with the trend sound on it. This guide goes deeper than the trend page: how the skit is built, where it came from and what to expect from your own clip.
Key takeaways
- The earliest clip we found is dated 27 October 2026, so the trend is still emerging.
- Photo 1 is you; photo 2 is optional, and an empty box gives you a made-up TV-star type, not a real celebrity.
- Photos of celebrities, public figures and children are not accepted.
- A 15-second Standard 720p clip is 13 credits, a first video costs no more than 10, and the Single pack is $6.99 for 10 credits.
- The original sound is on the video, but nobody lip-syncs.
In this guide
- What is actually happening in the 15-second skit?
- How do the six beats line up on the timeline?
- Where did the trend start, and how young is it?
- Why do people call it the celebrity kitchen, and what are the other names?
- Which AI model makes the clip, and why does it need a motion reference?
- What happens to the second person if I leave the box empty?
- Can I upload a photo of a celebrity?
- Which photo gives the best likeness?
- What about the sound, and does anyone talk?
- Do I need to label the clip as AI when I post it?
- What does a celebrity kitchen clip cost, and what if it fails?
What is actually happening in the 15-second skit?
Two adults stand at a speckled granite counter with a red mug, a jar of instant coffee, a kettle and a jug of milk. The one in front spoons coffee onto the counter beside the mug, straight-faced, again and again. The one behind fills the empty mug with hot water from the kettle, then grabs the milk and pours it straight onto the counter in a big white splash.
The joke is the lack of any reaction until the very end. Nobody explains the mistake, the camera sits low at counter level and never moves, and the last second is two people staring at the mess. That restraint is why the clip is easy to watch on a loop and easy to copy.
How do the six beats line up on the timeline?
Roughly like this: 0 to 2 seconds is the first spoonful landing beside the mug; 2 to 4.5 seconds is the kettle filling the empty mug; 4.5 to 9 seconds is spoonful after spoonful on the counter; 9 to 12.5 seconds is the grin as the milk jug is lifted; 12.5 to 13.5 seconds is the spill; 13.5 to 15 seconds is the raised hands and the stare.
If you plan to post your own version, the grin at 9 seconds and the spill at 12.5 seconds are the two moments viewers wait for. Both are built into the template, so you do not have to direct them.
Where did the trend start, and how young is it?
The earliest clip we found is an AI fan edit posted on TikTok on 27 October 2026. CapCut templates and Instagram and YouTube tutorials that call it the celebrity kitchen followed on 28 October. We could not confirm whether an older live-action original exists, so treat the origin as an AI-era meme rather than a remade sitcom scene.
The page was added on 29 October 2026, which makes this an emerging trend. Emerging means the format is still being copied heavily and nobody has run out of angles yet, but it also means details such as the sound and the most popular variants can change within days.
Why do people call it the celebrity kitchen, and what are the other names?
The same clip travels under several labels: celebrity kitchen on Instagram and YouTube, trying to make coffee, coffee fail and milk spill in search and in template names. They all point at the same skit. If you are searching for a template or a tutorial, try two or three of those phrases, because creators rarely agree on one.
Naming also tells you what viewers expect. A post titled for the coffee fail sets up the gag; a post titled celebrity kitchen sets up the idea of famous faces in an ordinary room. Your caption can use either, but the clip underneath is identical.
Which AI model makes the clip, and why does it need a motion reference?
videosi renders it with ByteDance's Seedance model, version 2.0 or 2.5 depending on the tier you pick. ByteDance describes Seedance 2.0 on its official page as a multimodal audio-video model that accepts text, image, audio and video inputs, with images and video usable as references for performance and camera movement (ByteDance Seed, Seedance 2.0).
That reference input is the whole trick. A plain text-to-video model has to invent the timing, so the coffee often lands in the mug or the milk never spills. With the original 15-second clip as a motion reference and your photo as the identity source, the kitchen, the props, the camera and the order of beats stay put while the people change.
What happens to the second person if I leave the box empty?
Photo 1 is you at the counter. Photo 2 is optional and becomes the person with the kettle and the milk. If you leave it empty, videosi casts a made-up adult with the polished look of a TV star, in a dark jacket, who is not meant to resemble any real celebrity.
That is a deliberate choice. It lets the clip read as a celebrity kitchen, with you as the famous one, without copying a real person's face. If you add a partner or friend, use one adult per photo and only people who agreed to appear. Adding a co-star costs nothing extra.
Can I upload a photo of a celebrity?
No. Photos of celebrities and public figures are not accepted, and neither are photos of children. The trend name can make this sound like the point of the exercise, but the product only puts the people in your own photos into the kitchen.
There is a practical side too. A face you do not own rights to is a poor starting point for a clip you want to post, and platforms are increasingly strict about synthetic media that could mislead. The made-up co-star is the safe way to get the celebrity feel.
Which photo gives the best likeness?
Use a front-facing, evenly lit photo of one person with the whole head and shoulders in frame. Avoid sunglasses, masks, heavy filters and group shots, since each of them gives the model less to hold on to. If your outfit matters, wear it in the photo, because the identity lock covers face, hair, build and outfit.
Be realistic about the result. Likeness is not guaranteed, and the face drifts most in the moments when you look down at the coffee or turn to your co-star. If the first render is off, change the photo before you change anything else.
What about the sound, and does anyone talk?
The clip carries the original TikTok sound labelled "original sound - NS7", posted with the earliest clips and reused by the template posts. No named song is confirmed, so there is nothing to license or look up. videosi puts the same sound on your video automatically.
Nobody lip-syncs. Mouths move loosely or stay closed, which suits a skit built on silence and facial expressions. If you want spoken lines, add them in your editor after downloading.
Do I need to label the clip as AI when I post it?
Label it. Most major platforms have rules or tools for marking AI-generated content, and a realistic video of you and a famous-looking co-star in a kitchen is exactly the kind of clip that can be mistaken for real. Check the current rules of the app you post on, since they change.
If you post it as part of a promotion, the US Federal Trade Commission's guidance is also relevant: it says material connections should be disclosed clearly, near the endorsement and, in video, both visually and audibly (FTC, Endorsement Guides: What People Are Asking). A joke clip with no brand involved is a simpler case, but an honest AI note costs nothing. videosi is independent and not affiliated with TikTok, CapCut or any actor, show or brand in the original clips.
What does a celebrity kitchen clip cost, and what if it fails?
A 15-second clip on the Standard model at 720p costs 13 credits, and an account's first video costs no more than 10. The Single pack is $6.99 for 10 credits, so a first video fits inside one Single pack. Credits never expire, and a failed, timed-out or blocked render returns its credits automatically.
Rendering takes several minutes, so start it and come back. You can pick 9:16, 1:1 or 16:9 before you press Generate. The full flow, with the sample and examples, is on the celebrity kitchen AI video page.
Ready to try it? Open the celebrity kitchen AI video page, upload one photo and press Generate.
