The Coffee Cat AI trend is a 15-second clip in which two cats sip iced coffee at a café table and one of them stands up and dances, and it is made by AI video models, not filmed. On videosi you get the same clip starring your own pet from one photo, with the trend sound already on it and no prompt to write. This guide answers the questions people ask after the basics.
Key takeaways
- The earliest iced-coffee copy we could date is from 19 October 2026 on TikTok, so the trend is emerging, not peaked.
- There is no single confirmed song, and no original creator has been identified.
- One photo of your pet is enough; a second photo for the friend beside it is optional and free.
- A 15-second Standard 720p video is 13 credits, and failed renders return their credits.
- Label the post as AI-generated before you publish.
In this guide
- What is the Coffee Cat trend, exactly?
- Where and when did it start?
- How do I make a Coffee Cat video of my own pet?
- Which photo works best?
- Does it work for dogs and other pets?
- What happens if I do not add a second pet?
- What song is used?
- Which AI model makes it, and can I use other tools?
- How much does it cost?
- Why did mine come out wrong?
- What should I know before posting it?
What is the Coffee Cat trend, exactly?
It is a 15-second AI clip shot like a fixed café camera. Two cats sit side by side at a marble table at night, each with a clear iced coffee and a red straw. They sip, stare into the lens and swat at each other. Then one rises onto its hind legs and dances with its arms waving while the other slumps behind its cup, and at the end the dancer drops out of sight below the table edge. Nothing in it is filmed: the cats, the cups and the terrace are all generated.
The humour is the caffeine hit. Posts caption it as the friend who was not ready for iced coffee at 3 a.m., and the dance is the punchline. The whole thing is a repeatable template, which is why it can be remade with other animals.
Where and when did it start?
We found it on TikTok in October 2026. The earliest iced-coffee version we could date is from 19 October 2026, and a second one is from 23 October. Each uses a different pair of cats and a different song, and one of them sits on a beach terrace with black umbrellas instead of a night terrace with a lit ferris wheel. A Reels version of the same clip also circulates.
We could not identify who posted the first one, so the page on videosi lists the origin as unconfirmed. The wider family of dancing-cat clips is much older: videos in the millions of views have been around since August 2026. Coffee Cat is a small, recent variation inside that big trend, and as of 1 November 2026 most iced-coffee copies have only hundreds of views.
How do I make a Coffee Cat video of my own pet?
Open the Coffee Cat AI page on videosi, upload one clear photo of your cat or dog, and press Generate. You pick the model tier, the resolution and the format (9:16 or 16:9), and a few minutes later you get a 15-second MP4. You do not write a prompt.
The photo is for the dancer, the pet that stands up in the clip. A second photo is optional and becomes the friend sitting beside it. Without it, videosi adds an ordinary made-up cat, so a single photo is always enough.
Which photo works best?
Use a photo of the whole animal, sitting up or standing, taken at its eye level in daylight. The dancer has to stand on its hind legs for several seconds, and a face close-up gives the model nothing to build that body from, so it invents one and the likeness suffers.
Keep one animal per photo. If you want a real pair, send one photo with only the dancer and a second photo with only the friend. Two pets in one frame can merge or swap places. A dark pet in a dark room also loses the markings that make it recognisable, so a bright window is better than a lamp.
Does it work for dogs and other pets?
Yes. The clip is built around cats, but a photo of a dog is placed in the dancer's seat and the dog sits up and dances the way the cat does. Expect the moves to look less natural on a dog, because the arm-waving dance comes from the cat in the original.
You can also use a friend's pet or a rescue you are fostering. Only use animals you have the right to show, and keep in mind that the clip never includes people, so there is nothing to upload of yourself.
What happens if I do not add a second pet?
Nothing breaks. The friend at the table becomes a different ordinary house cat that videosi makes up, with a coat that does not match the dancer. Adding the second photo costs nothing extra, so the only reason to skip it is that you do not have one.
If both pets are in one photo, the generator also takes a single group photo: the dancer on the left, the friend on the right. That is quicker, but two separate photos give a more reliable likeness.
What song is used?
There is no single confirmed song. The two TikTok copies we checked use different tracks, and the music in the original clip is not identified, so we do not name one. The video you make on videosi comes with the audio of the reference version already on it, so it does not need CapCut to sound right.
Platforms can mute audio they do not recognise. If yours is muted, replace it with a sound from the platform's own library, which is also the safe choice for copyright.
Which AI model makes it, and can I use other tools?
videosi runs ByteDance's Seedance 2.0 and 2.5, with a motion reference clip, and you choose the tier. The prompt is not public, but in short it tells the model to replace the dancing cat with your pet and keep the table, the cups and the 15-second timing.
You can try other video tools, but a plain text prompt rarely gets the timing right. The dance, the sip and the drop out of sight need to line up inside 15 seconds, and two animals with props make that hard. A motion reference is what keeps the beats in place.
How much does it cost?
A 15-second Coffee Cat video on the Standard model at 720p is 13 credits. Better models and higher resolutions cost more. There is no free generation. Credit packs start at $6.99 for 10 credits and never expire, subscriptions start at $29.90 a month, and a failed render returns its credits.
Why did mine come out wrong?
The usual causes are a face close-up photo, a very long coat that drifts, rare markings, or a dark photo. The dancer also changes the most while it is standing and dancing, so check that moment first. If the friend looks like a copy of the dancer, you gave one photo for both, so add a second photo of a different animal.
Likeness is never guaranteed, and the page says so. Try again with a brighter, whole-body photo before changing anything else.
What should I know before posting it?
Label it as AI. TikTok's community guidelines say "we require creators to label AI-generated or significantly edited content that shows realistic-looking scenes or people." A café cat that looks filmed is the kind of realistic scene it describes, so turn on the AI-generated label and say so in the caption.
Do not credit a first creator by name, because that is unconfirmed. When you are ready, open the Coffee Cat AI trend page and make yours.
