Macarena AI is a short video in which a person, a group of friends or a pet from one photo performs the 1990s Macarena dance to the original song. As of November 2026 the trend is past its peak and fading, but the dance and the song stay recognisable, so the format still works for a quick, funny clip.
Macarena AI on videosi turns one photo into a 14-second video with the moves, the timing and the song already built in. This guide goes past the basics and answers the questions people ask once they have seen the trend and want to know how it works, why it looks the way it does and what to expect.
Key takeaways
- The song behind the trend is "Macarena" by Los del Río, recorded in 1993, and its Bayside Boys remix topped the Billboard Hot 100 for 14 weeks in 1996.
- The AI wave is past its peak as of November 2026, so treat it as a fun throwback rather than a rising trend.
- A photo plus a motion reference is what keeps the arm moves in the right order; a text prompt alone usually invents a generic dance.
- Full-body, front-facing photos of adults who agreed (or of a pet) give the best likeness. Children's photos cannot be uploaded and there is no baby version.
- The finished MP4 keeps the original song and has no watermark. A 14-second video costs 12 credits at Standard 720p, 10 for an account's first one.
In this guide
- Why does the Macarena suit AI video so well?
- What song is used, and what is its history?
- Where did the Macarena AI trend actually start?
- Why does a motion reference matter more than the prompt?
- How can I get the best likeness from my photo?
- What can the three settings change, and what stays fixed?
- Can two or three people, or a pet, do the Macarena together?
- Why can't I upload a child's photo, and why is there no baby version?
- Is the sound on the finished video, and is there lip sync?
- What does it cost, and what happens if a render fails?
- Is it worth making a Macarena AI video now that the trend has faded?
Why does the Macarena suit AI video so well?
The Macarena is one of the few dances that everyone can picture move by move. It is a fixed sequence of arm and hip gestures done in place, facing the camera, so a viewer knows at once whether the video got it right. That makes the joke clear: a still photo, or a person who plainly cannot dance, suddenly runs the whole routine without a mistake.
It also suits the technology. The dancer stays in one spot, the camera does not move and the clip is a single continuous take. Those are exactly the conditions where video models hold a face and a body together best, compared with a spinning dance or a fast camera move. The trade-off is that hands cross, fold and go behind the head, which is where AI video tends to slip.
What song is used, and what is its history?
The song is "Macarena" by the Spanish duo Los del Río, Antonio Romero Monge and Rafael Ruiz Perdigones. They first released it in 1993. According to a Grammy.com feature, a Miami radio personality, Johnny Caride, noticed it in local clubs in 1995 and had the Bayside Boys record a remix with English lyrics, which RCA later acquired.
That Bayside Boys remix reached No. 1 on the Billboard Hot 100 in August 1996 and held the top spot for 14 weeks. Stereogum's Number Ones column adds that the original, non-remixed version peaked at No. 23 in the US, so the remix is the version most people remember. Sources disagree on some details such as total weeks on the chart, so only the 14 weeks at No. 1 is quoted here.
Where did the Macarena AI trend actually start?
No single original creator is confirmed. The earliest filter-style post we watched is from late December 2025, and an anime AI Macarena dance from July 2023 predates it. In December 2025 and January 2026 posts presented the dance as a filter that makes you and your friends better dancers, and a large wave of CapCut templates followed in September and October 2026.
Because the idea is old and many people remade it at once, it behaves less like a single viral post and more like a format that keeps resurfacing. That is also why the exact name of the filter changes from app to app.
Why does a motion reference matter more than the prompt?
A text prompt can say "dance the Macarena", but a video model has to guess what that means. Without a reference it tends to produce a generic party dance and lose the exact order of the moves: arms out with palms down, palms up, hands to opposite shoulders, hands behind the head, hands on the hips, the wiggle and the final kick.
videosi gives the model a 14-second reference clip that already contains that order, the rhythm, the still full-body camera and the framing. The photo then supplies who is dancing. Splitting the job this way, motion from the clip and identity from the photo, is why the routine comes out in sequence instead of as a vague sway. The full prompt is not published.
How can I get the best likeness from my photo?
The clip is a full-body shot, so your face is small in the frame. Use a sharp, front-facing photo in even light that shows your whole body from head to shoes. Leave off sunglasses and low hats, and avoid heavy shadows across the face. Your hair and clothes in the photo are the ones you dance in, so pick an outfit you like.
Even with a good photo, fast arm moves can blur a hand or make a face drift, so likeness is not guaranteed. A waist-up crop is the most common cause of odd results, because the model has to invent your legs and shoes. If the first result is off, try a cleaner photo before changing anything else.
What can the three settings change, and what stays fixed?
You can keep the warm panelled room from the reference clip, move to a bright dance studio or move to a paved city promenade at night. Whatever you pick, all 14 seconds stay in that one place, because the clip has no cuts.
What does not change is the choreography, the timing, the camera position and the song. Your own photo background is not reused, so if the place matters, choose the closest of the three. The format is also your choice: 9:16 for TikTok, Reels and Shorts, 1:1 or 16:9.
Can two or three people, or a pet, do the Macarena together?
Yes, within limits. One photo can show up to three adults, standing side by side, each keeping their own face and clothes and doing the same moves in time. Choose a photo where everyone is fully visible and nobody overlaps. More people means smaller faces, so likeness gets less reliable as the group grows.
A pet is also possible: one animal, whole body, on a plain background works best. The pet stands upright on its hind legs and does the moves with its front paws while staying a real animal. Fur patterns carry over better than unusual body shapes, and pets vary the most from photo to photo.
Why can't I upload a child's photo, and why is there no baby version?
Baby Macarena clips are popular in CapCut templates, but videosi's Macarena AI page does not make them. The page is for adults and pets. Do not upload photos of children, and only use photos of adults who have agreed. The same goes for celebrities and strangers: use your own photo or one a friend has said yes to.
This is a deliberate limit on the page, not a technical gap you can work around. If you want the dance with family, use a photo of the adults in the family or a pet, which gives the same routine and the same song.
Is the sound on the finished video, and is there lip sync?
The finished MP4 carries the original Macarena song, taken from the same 14-second section as the motion reference, so it is ready to post. There is no lip sync, because the dance is the point and nobody needs to mouth the words.
The video has no watermark. You should still label it as AI-generated when you post, in line with what each platform asks. Rendering takes several minutes, and the result is saved in My creations so you can download it later.
What does it cost, and what happens if a render fails?
A 14-second Macarena AI video costs 12 credits at Standard 720p, and an account's first video at that setting costs 10. It is not free and there is no free trial generation. The smallest pack is $6.99 for 10 credits, credits never expire, and plans start at $29.90 a month.
If a render fails, times out or is blocked, the credits are returned automatically. You do not write a prompt, so there is nothing to refine between tries except the photo, the setting and the format.
Is it worth making a Macarena AI video now that the trend has faded?
As of November 2026 the AI wave is past its peak, though new posts, including Halloween versions, were still appearing in late October. That means a Macarena clip will no longer ride a surge of attention, but it also costs little to make and needs no explanation, since the song and the dance are known across generations.
It works best as a personal or group joke: a birthday message, a party invite or a reply in a chat. If your goal is reach on a trending sound, expect less than the September and October peak. The song itself keeps returning, which is why the dance stays recognisable even as this wave cools.
Ready to try it? Open the Macarena AI page, upload one photo, pick a setting and press Generate.
