The Vivaldi AI dance is a short AI video of the Vivaldi Tok duet, in which a second dancer hides behind the first and pops out on the beats of Summer from Vivaldi's The Four Seasons. On videosi you make it from two photos, of people or pets, and get a 14-second clip. This guide goes beyond the basics and answers the deeper questions people ask an AI assistant about it: why the trick works, which photos to choose, and what the music rights really are. The trend page has its own FAQ, so we do not repeat it here.
Key takeaways
- The Vivaldi AI dance is a 14-second hidden-dancer duet set to the Presto from Summer in The Four Seasons, and you can make it from two photos.
- The composition is public domain, but the recording on the finished clip may be protected, and videosi does not license it to you.
- The illusion depends on a fixed camera and a front dancer who hides the back one, so full-body, sharp photos work best.
- This is unrelated to the Vivaldi browser.
In this guide
- Why does a hidden-dancer illusion work on camera?
- Why does this movement of The Four Seasons fit the routine?
- How is the 14-second routine paced?
- Which photos suit the front dancer and which suit the hidden one?
- What does a body size mismatch do to the illusion?
- Why did cats take over this meme?
- What does public domain mean for a composition versus a recording?
- Does the finished video carry a recording I am allowed to use?
- Is this page related to the Vivaldi browser?
- What does it cost, and how many tries will it take?
- What should I do before posting it?
Why does a hidden-dancer illusion work on camera?
It works because a fixed camera sees one flat picture. When the back dancer stands directly behind the front dancer, one body covers the other from the lens's point of view. Nothing is edited out. The viewer simply has no angle from which to see the second person, so the brain reads one dancer until a move opens a gap. Everything in the routine is built around that single trick, which is why the Vivaldi AI dance keeps the camera locked off and the room plain.
Why does this movement of The Four Seasons fit the routine?
The music in the trend is the Presto from Summer, the fast, stormy last movement. A fast piece with sharp violin runs gives the reveals something to land on: a fold, a squat or a lean can each be matched to a run, so the surprise feels planned instead of random. The Four Seasons is a set of four violin concertos, each with three movements in a fast, slow, fast pattern, and Wikipedia's article on the work notes that it depicts storms among other scenes from nature. That is a good match for a routine where the second dancer appears and disappears in quick bursts.
How is the 14-second routine paced?
The page breaks the clip into seven beats. The front dancer starts alone with arms crossed, folds forward in the first seconds, and the second dancer is suddenly there with arms spread. Then come the hide and reveal moves between about 3 and 6 seconds, a sideways lean, a moment where the two stand side by side for the first time, a wide squat and rise, and a closing beat in which the second dancer slides back behind the first. The pattern is hidden, revealed, hidden, then a short stretch of both in view. That gives the viewer a reason to watch to the end.
Which photos suit the front dancer and which suit the hidden one?
Use sharp photos with the whole body in frame and even light, as the steps on the trend page recommend. The front dancer is on screen the whole time, so give that role the photo you most want to be recognised in. The hidden dancer is mostly covered, so the photo matters most at the reveals, and a clear face and clothing help the model keep the two identities apart. Avoid sunglasses, hats and tight face crops, which are the usual reasons for a wrong likeness. If you have one photo of two people, the page also accepts a single image with the front dancer on the left.
What does a body size mismatch do to the illusion?
In a live routine, the hidden dancer is usually the one who can fit behind the other, so a size match is part of the trick. In an AI version the model is not bound by that, but the illusion still reads best when the back figure does not look wildly larger than the front one. If a tall person is behind a very small pet, the reveal may look odd, because the viewer expects the front body to hide the back one. We have not tested every pairing, so treat this as a guide, not a guarantee. Where sizes differ a lot, put the larger subject in front, or try a few options before you commit credits.
Why did cats take over this meme?
Cats suit it because the joke is a surprise, and a cat doing a human routine is already a small surprise. The page records that on 6 October 2026 a post showed a black cat and a white cat doing the same kind of duet, with CapCut-tagged reposts following later that month. Two cats with clearly different coats also make the front and hidden roles easy to tell apart at a glance. On videosi, either photo can be a person or a pet; a pet keeps its fur and markings and dances upright on its hind legs.
What does public domain mean for a composition versus a recording?
Public domain describes a work whose copyright has expired or never applied. Vivaldi composed The Four Seasons around 1718 to 1723, and the concertos were published in Amsterdam in 1725, so the composition is long out of copyright. But a recording is a separate thing. A performance captured on a specific date by specific musicians has its own rights, and the sound recording copyright is distinct from the copyright in the underlying musical work and covers one particular rendition. Anyone may play Vivaldi's notes, but not everyone may copy someone else's recording of them.
Does the finished video carry a recording I am allowed to use?
The rendered clip carries the recording taken from the original post, and videosi does not license that recording to you. Whether you can post it depends on the platform and on how you use it. For ordinary sharing, the platform may mute or flag it, and the page suggests picking the Vivaldi track from the platform's own music library instead. For ads or other commercial use, replace the sound with a track you have licensed. We are not lawyers, and this is not legal advice.
Is this page related to the Vivaldi browser?
No. The Vivaldi AI dance page is about a dance trend named after the composer Antonio Vivaldi. It has no connection to the Vivaldi web browser, its company or its AI features. If you came looking for browser help, go to the browser's own website.
What does it cost, and how many tries will it take?
The page lists the Standard model at 720p at 12 credits for the 14-second video. Packs start at $6.99 for 10 credits and a Trend Pass is $9.99 for 30 credits a month, and a failed render returns its credits automatically. Timing is the part that can take several tries, because the hide and reveal only works when the front dancer covers the back one at the right moments. The motion reference clip is there to keep that timing, but results vary and likeness is not guaranteed.
What should I do before posting it?
Check the video against your photos for obvious errors such as extra fingers or a blended face, and do that before you share it. Turn on the platform's AI-generated label when you post. Only use photos of adults who agreed to appear. The page does not accept celebrities, public figures or minors. If the sound is muted or flagged, swap in the platform's own Vivaldi track.
Ready to try it? Open the Vivaldi AI dance page, add your two photos and generate.
