Horse-riding dance · AI remakes since October 12, 2026
Gangnam Style AI Dance: Set the Scene, Then Trot
Upload your photo, pick or describe a set, press Design the frames, then Animate it to land on the lasso pose.
Last checked 2026-10-29
⚙ AI Forge Studio
Pick or describe the set, design the first and last frame, then animate the trot.
1First frame · 3→
2Last frame · 3→
3Animate 10 s · 280
Your photo
9:16 vertical · first frame, then last frame, then the clip between them
Next: last frame 3 credits, then animate 10 s · 280 credits.
Joining the trend is free; making it on Lovart ME uses credits from packs starting at $9.90 for 250 credits. Each button shows its exact cost before you generate.
SAMPLE · INPUT → RESULT
Sample inputs and result from this page. Your result will appear here.
Start with an idea
01 / Samples
Gangnam Style AI dance samples cut for Lovart
A crosswalk pair and a rooftop solo remixed from 10-second originals, plus an elevator set still designed here.
The horse-riding dance hops from foot to foot like a trot with the wrists crossed as if holding reins, then one arm swings an invisible lasso over the head. Every dancer below is an original generated adult; no real dancer’s face, music video frame or original recording is used. The two clips are Lovart ME cuts of Seedance 2.0 Fast originals; the Elevator set still was made with GPT Image 2.5 Flare on October 29, 2026. None of them came from the workbench above.
Trot at the crosswalk
Photo 1Photo 2→Still→StartEnd
Trot at the crosswalk
Remixed from a Seedance 2.0 Fast clip
Inputs
Two original portraits: a man in his late 20s with a buzz cut and a mint varsity jacket (left) and a woman in her late 20s with a long black ponytail and an oversized white hoodie (right)
Set still
GPT Image 2.5 Flare: both full body on a night crosswalk lined with neon shop signs, wrists crossed as if holding reins
Original
Seedance 2.0 Fast · 9:16 · 10 s, the set still as the only first frame, made October 29, 2026; original electronic beat from the prompt
In the clip
Both trot with crossed wrists, swing one arm over the head at about 2 s, trot again on the wet stripes and finish pointing at the lens
This cut
8.8 s, re-cut, reframed and steel-blue graded; no new AI generation
One original portrait: a woman in her 30s with a short black bob, a red track jacket and black trousers
Set still
GPT Image 2.5 Flare: a city rooftop at dusk with water tanks and a pink sky, wide stance, wrists crossed
Original
Seedance 2.0 Fast · 9:16 · 10 s, made October 29, 2026
In the clip
Crossed-wrist hops from foot to foot, one arm swung out to the side at about 3 s, knee hops, then a grin; the arm never goes fully overhead in this cut
This cut
8.4 s, re-cut, reframed and graded; no new AI generation
The woman’s portrait from Trot at the crosswalk, plus the text “neon elevator”
Still
GPT Image 2.5 Flare image to image, 1024 × 1536, made October 29, 2026: mirrored steel walls, pink and cyan light strips, her wrists stacked at the chest, knees bent
Output
Still only: a first frame for the Neon elevator set; it has not been animated
Upload your photo plus up to two set pictures, describe the place, and Image to Image builds a first frame of you standing there.
Your photo
One person per photo, face to the camera, full body or waist up with space above the head for the lasso arm. For Night crosswalk add a second person; Swap left and right fixes who stands where.
Set pictures
Optional: up to two photos of a real place, your office, a stairwell or your own roof. They only steer the walls, light and colors. Describe the place in a few words, like “neon elevator”, and pick a set chip to start from.
The first frame
Design the frames sends your photo, the set pictures and the place to Image to Image (GPT Image 2.5 Flare by default) and returns you standing there, wrists crossed, ready to trot. Check the face before you go on.
Neon elevator
Mirrored steel walls and two colors of light strip. A tight space, so the trot reads as a joke; keep the frame full body.
Night crosswalk
Wet stripes and neon signs for two people side by side. Photo 1 stays left, Photo 2 stays right for the whole clip.
Rooftop at dusk
An open sky above the head leaves room for the lasso swing. Good for one dancer.
Make a last frame with one arm high mid-lasso and put it in the end slot, so the clip stops on the move everyone knows.
First frame · wrists crossed→The lasso pose · clip frame at 2.3 s
The right image is a frame from the Trot at the crosswalk clip at 2.3 seconds, shown as a guide to the pose. That sample was animated from a first frame only; no designed last frame was used for it.
First frame
Wrists crossed in front of the chest, knees bent, weight ready to hop. This is the frame Design the frames returns, standing in your set.
Last frame
Design the last frame edits that same picture: same set, same framing, same clothes, with one arm raised high as if spinning a lasso and the other hand at the hip. Animate it then runs Image to Video between the two frames, 10 seconds with Seedance 2.5 by default; the cost shows on each button.
Take the last frame into PhotoCraft, add a title and your names, and post the poster next to the clip.
Start from any designed frame; the Elevator set still is shown here before type.
Type in PhotoCraft
Download the last frame and open it in PhotoCraft, the free browser editor. Put a short title across the top, your names and the date along the bottom, and keep the type clear of the raised arm.
Or the Poster chip
Pick Poster in the Gangnam Set tab for a still only: your photo in the set you describe, one arm high, with HORSE DANCE NIGHT set across the top. Check the spelling before you print; the type is drawn by the image model.
Why the Gangnam Style AI dance is everywhere this month
An October 5, 2026 selfie reel from Seoul's Gangnam district went viral, AI remakes began October 12 and film-character versions followed on October 19.
Oct 5, 2026A selfie reel filmed at night in Seoul’s Gangnam district, three friends taking turns at one camera, goes viral: 17.5 million views by October 22, per peer pages.
Oct 12, 2026The first AI remake: three AI-made women do the dance at sunset.
Oct 19, 2026Film characters doing the horse-riding dance pass 600,000 views.
Oct 21, 2026Phone-editor template tutorials follow, and generator pages go up.
Post counts
#gangnamstyleai went from 255 videos on October 22 to 566 on October 23, 2026; #gangnamstyletrend had 773, per peer pages.
The sound
Generators return the clip without the original recording, with an original beat or none. People add the Gangnam Style sound in TikTok or Reels when they post.
Yes. Type the place into Describe the place, for example “glass office lobby”, or pick Neon elevator. Add up to two pictures of the real room as set pictures and the first frame takes its walls and light; the set pictures are never used for a face.
Why use a last frame?
The last frame shows you mid-lasso with one arm high. With a first and a last frame, the video model has to arrive at that pose, so the clip stops on the move people recognise instead of a random step halfway through the trot.
Where did the film-character versions come from?
Peer trend pages report that a TikTok account posted film characters doing the horse-riding dance on October 19, 2026, a week after the first AI remake on October 12.
Is the song included?
No. The clip comes with an original beat or no music. Add the Gangnam Style sound from the TikTok or Reels sound picker when you post.
What photo works for the frames?
Full body or waist up, face to the camera, evenly lit, with space above the head so the lasso arm is not cut off. One person per photo; no sunglasses, masks or beauty filters, and only adults who agreed to it.
Where do the sample clips come from?
They are Lovart ME cuts of 10-second Seedance 2.0 Fast clips made for this trend on October 29, 2026 from the set stills on the cards: re-cut, reframed and color graded, with no new AI generation. The Elevator set still was made on this page’s sample pipeline, not in the workbench.
Lovart ME is an independent site and is not affiliated with the song’s artist or label, TikTok, MakeViral or Kavel. This page copies only the dance moves: no original recording, music video footage or original dancers’ faces. Sample people are original generated adults.
