The kung fu cat AI video is a 15-second clip in which a cat sits upright on a lap and chops at swinging paper towel to Kung Fu Fighting while keeping a completely blank face. On videosi you make it from one photo of your cat, with the song already on the file and no filming at all. This guide goes past the basics: why the stare works, which prop to pick, how to choose the photo, and what to do before you post.
Key takeaways
- The joke is the contrast between fast paws and a blank stare, so choose a photo with a calm, clear face.
- Paper towel is the classic prop; the banana is a variation we have not test-rendered.
- Other tools tend to invent the rhythm and drop the song, which is why a reference clip matters.
- Making it from a photo means your cat is never handled; label the post as AI before you publish.
- A standard 15-second video is 13 credits, and failed renders return them.
In this guide
- Why is a blank stare funnier than an angry cat?
- What is the difference between the paper towel and banana versions?
- Which photo of my cat should I pick?
- Why does Kung Fu Fighting make the joke work?
- Why can other AI tools not hit the beat?
- Is the AI version kinder to my cat than filming it?
- If I film it myself, how do I keep my cat safe and comfortable?
- Why do the paws look odd in the fastest chops?
- How do I label the post as AI?
- What does it cost, and what is still untested?
Why is a blank stare funnier than an angry cat?
Because the joke is a gap between what the paws do and what the face does. A cat that looks angry reads as a cat fighting. A cat that chops at full speed while staring at the lens like it is waiting for a bus reads as a tiny master who finds the whole thing beneath them. The page's beat table shows how it is built: a guard pose, a first strike, a reset, a second round, low kicks, then a long frozen stare and one raised back paw.
Those pauses matter as much as the chops. Every time the paper disappears, the cat goes back to the same neutral face, so the audience gets the contrast again and again. When you make your own, the stare is the part to protect. A photo where your cat looks calm and straight ahead gives the model the right face to keep.
What is the difference between the paper towel and banana versions?
The paper towel is the original prop. It is a thin white sheet that a second person swings, so it flutters, folds and drops, and the chops land on something that visibly reacts. The banana is the fruit variant of the trend, which also shows up with pineapple, kiwi, apple and cucumber in the clips we looked at.
On the kung fu cat AI page you pick one of the two before you generate. The paper towel keeps the reference clip closest to how it was filmed, because the sheet stays in the same place and swings on the same beats. The banana replaces the sheet, so it has to behave like a solid object in the same spot, and it may look less natural. We have not test-rendered either option for the page, so treat the banana as the more experimental choice.
Which photo of my cat should I pick?
Pick one sharp photo of one cat, with the face and both ears clear and the front paws in view if you have it. A cat sitting upright is ideal, because the whole clip is an upright cat on a lap. Soft window light shows fur detail better than a flash.
Two other choices help. First, keep the cat as it normally looks, because hats, costumes and blankets can confuse the swap. Second, favour a photo where the markings are visible: stripes, a white chest, socks or odd-coloured eyes are what make a viewer say "that is my cat". If the photo is blurry or hides the face, the model falls back on the ginger tabby in its reference clip, and you get a cat that does not look like yours.
Why does Kung Fu Fighting make the joke work?
The clip is built around the sound. The trend uses the opening of Kung Fu Fighting, written and recorded by Carl Douglas and produced by Biddu. It came out in August 1974, reached number one in the UK in September 1974 and also topped the Billboard Hot 100, according to Wikipedia. Everybody was kung fu fighting is the line most people know, and it is the one a cat on a lap is acting out.
That is the reason a silent version would be a worse joke. The chops are cut to the beat, and the sound tells the viewer what the cat is meant to be doing before they see a single paw. Every video from videosi comes with the trend's original sound already on it, so you do not need to find the song or sync it. On TikTok the song is often attached as a creator's original sound, so the label you see may differ from the title.
Why can other AI tools not hit the beat?
Most general video tools take a text description and invent the rhythm. A description such as a cat chopping paper towel to a disco beat gives you a cat in a martial arts pose, often standing on a table, and the chops are not tied to anything. The timing of the original is a sequence of specific moments, and words do not carry them well.
Two other things usually go wrong. The first is identity: if the instructions only borrow motion from a reference, the original ginger cat can stay in the frame with a few details from your photo. The second is sound, because many generators return silent video, and a kung fu cat with no song loses most of its timing.
videosi handles this by using a 15-second reference clip for the motion and your photo for the animal, with the song put back on top. It runs on Seedance from ByteDance, and the prompt follows a fixed structure: swap the animal, lock its identity, set the prop, follow the timing, and add rules such as one animal only and no text. We do not publish the full wording. If you try another tool, look for one that accepts a reference video, and expect several attempts.
Is the AI version kinder to my cat than filming it?
Yes, in the sense that nothing happens to your cat at all. You upload one existing photo, and the video is drawn from it. Nobody has to hold your cat upright, swing paper at its face or wait for it to react, and the cat can be asleep in another room.
In the filmed originals, a person's hands guide the paws, which asks a lot of an animal that has not chosen to do it.
If I film it myself, how do I keep my cat safe and comfortable?
You can, but keep it gentle and short. The ASPCA's cat care guide says to pick a cat up by placing "one hand behind the front legs and another under the hindquarters" and lifting gently, and adds "Never pick up a cat by the scruff of the neck or by the front legs." Reading that, the idea of holding a cat upright and moving its paws should make you pause.
Stop as soon as your cat wants to leave, and never swing anything close to its face. You can read the ASPCA's advice on their general cat care page.
Why do the paws look odd in the fastest chops?
Fast paws are where video models fail first. During the quickest chops a paw can pick up an extra toe, two paws can merge, or the fur can smear for a frame or two. The reference clip is fast by design, so this is a known weak spot rather than a rare accident.
What helps is a photo where the front paws are visible, so the model has something real to hold on to, and a cat with a clear coat pattern. If the first result has a smeared moment, try a different photo rather than the same one again. Likeness is never guaranteed, and we have not test-rendered this page yet, so expect to try more than once.
How do I label the post as AI?
Switch on the AI-generated label in the app you post from. TikTok's integrity and authenticity guidelines say "we require creators to label AI-generated or significantly edited content that shows realistic-looking scenes or people," and add that unlabeled content may be removed, restricted or labeled by TikTok, depending on the harm it could cause. A cat on a lap looks like a filmed scene, so it is exactly the kind of clip to label. You can read the rule on TikTok's community guidelines page.
A short line in the caption, such as "AI version of my cat", is a good second layer, and it also tells viewers why the paws are not guided by anyone. Do not credit a particular creator for the trend unless you know who made it, because the earliest poster is unconfirmed.
What does it cost, and what is still untested?
A standard 15-second video (Standard model, 720p) costs 13 credits on the kung fu cat AI page. Better models and higher resolutions cost more. Every generation uses credits. Credit packs start at $6.99 for 10 credits and never expire, plans start at $29.90 a month, and a failed, timed-out or blocked render returns its credits automatically.
The prompt has not been tested in real renders yet, so we cannot promise a perfect likeness or a perfect banana. The trend is also young: the most-watched post we found had 13.3 million views on 2 November 2026, and we have no figure for the trend as a whole. If it fits your cat, it is a quick thing to try while the format is fresh.
Want to see it with your own pet? Head to the kung fu cat AI trend page, or try the cat clapping meme next.
