Tutorial · updated October 15, 2026
How to make the Gangnam Style AI video, step by step
This tutorial shows how to make the Gangnam Style AI video that stays closest to the original clip: the same 21-second scene, the same cuts and camera moves and the same sound, with your faces in it. You need 1-3 photos and no prompt, editing app or download of the TikTok.
Switch between the original clip and a real result of this maker, unedited, made October 25, 2026 with photos of AI-generated sample people.
What you need
- 1-3 photos: You (the main one): opens and closes the video in the middle of the trio and gets two close-up solos. A person or a pet; Friend 2 (optional): gets the second solo, hair flip and all. Leave it empty and the AI adds a friend; Friend 3 (optional): gets the third solo by the glass lobby. Leave it empty and the AI adds a friend.
- 100 credits ($8.90 paid once, less with a plan), taken only when the video finishes.
- about 5–10 minutes of waiting. You can leave the page.
- A phone or a computer with a browser. No app, no editor, no account on another AI tool.
Step-by-step: make the Gangnam Style AI video
Open the template
Go to the Gangnam Style AI page on Remake Your Meme. The maker sits next to a real 21-second result, so you can compare it with the original clip before you spend anything.
Choose 1-3 photos
The Gangnam Style AI video needs 1-3 photos: photo 1 is you (the main one) (opens and closes the video in the middle of the trio and gets two close-up solos. A person or a pet); photo 2 is friend 2 (optional) (gets the second solo, hair flip and all. Leave it empty and the AI adds a friend); photo 3 is friend 3 (optional) (gets the third solo by the glass lobby. Leave it empty and the AI adds a friend). Use one person per photo with the whole face visible, ideally front-facing and sharp.
Upload and confirm
Drop the photos into the maker and confirm that you have permission to use them. They go straight to private storage and are deleted with the video after 30 days.
Keep the trend sound
Leave "Trend sound" on to get the audio of the original clip under your video. Choose "No song" if you want to add the official sound inside TikTok instead.
Start and wait
Press "Make my video". The Gangnam Style AI video usually takes about 5–10 minutes. You can close the tab or start another one; the video waits in Your videos, and we email you when it is ready.
Download and post
Download the 21-second 16:9 MP4 without a watermark and post it on TikTok, Reels or Shorts. Label it as AI-generated where the app asks, and tag the people in it.
Why this remake stays almost identical to the original
Most AI versions of this trend are generated from a prompt and a reference image, so the room, the timing and the camera drift away from the clip. Here the original clip is the starting point: we turned it into a depth video that keeps where everyone stands, how they move and how the camera moves, but drops the original faces. The video model films the scene again on that depth video, with empty sets rebuilt to match the original and your photos as the people.
Play the original and the remake together
Both clips start from the same second, so you can follow every shot and cut side by side.
What stays exactly as in the original
- The 21 seconds of the original night selfie in Gangnam, shot by shot: all three together, your solo, friend 2, friend 3, your second solo and all three again, with the fast whip pans between them.
- The handheld selfie camera and every head turn, hair flip and hand move, taken from the original clip as a depth video.
- The Gangnam street at night: office towers with lit windows, street lamps, passing cars and the Korean flags on their poles.
- Gangnam Style from the original clip, laid under the finished video, with everyone singing along.
What changes
- The three faces and outfits come from your photos; with one photo the AI adds two invented friends.
- A pet in any spot stays that pet: it sits up on its hind legs, bobs its head and opens its mouth with the song.
- The video is made in three parts and joined inside the whip pans, where the camera blurs, then sharpened to 1080p.
How to get the closest result
- Use sharp face photos: most of the video is close-ups.
- Night-out outfits (black tops, a white blouse) look most like the original.
- Use three photos so each solo shows someone you know.
Three ways to make a Gangnam Style AI video almost identical to the original
| Method | How close to the original | What you do |
|---|---|---|
| This maker | Very close: built from the original clip, with the scene, cuts, camera and sound already set up. | Add 1-3 photos and press one button. Failed videos are refunded. |
| Motion transfer, such as Higgsfield Genjutsu | Very close: Higgsfield says the motion, camera and timing stay as filmed and everything else is rebuilt from your references. | Find and download the original clip, upload it with your photos, describe the swap, check every shot, then add the sound back. |
| Prompt and a reference photo | Loose: the model invents the room, the timing and the camera again on every run. | Write a long prompt for every shot and retry until the scene looks right. |
Higgsfield's description read on its Genjutsu page on 15 October 2026. We are not affiliated with Higgsfield.
The manual way, for comparison
This is how people make the Gangnam Style AI video with a general AI video tool. It works, but every step is a place where the result drifts from the original:
- Find a Gangnam Style AI CapCut template or the Telegram bot from a tutorial and upload a photo.
- Hope the faces stay yours: many templates swap one face onto a fixed clip, and friend versions need three separate photos.
- Put the Gangnam Style sound under it in CapCut, then export it.
Photo tips for the Gangnam Style AI video
- Use one sharp photo per person, with only that person in it. The video has big close-ups, so a clear face matters more than the whole body.
- Everyone keeps the clothes from their photo. A night-out outfit looks most like the original.
- For a pet, use a clear photo of one animal with its face visible. It sits up on its hind legs and sings with the trio.
If you are looking for this video by description
People often ask for the Gangnam Style AI video without its name. These all mean the same clip, and this tutorial covers them:
- “the TikTok where three girls sing Gangnam Style at night in Korea”
- “gangnam style ai video with my friends”
- “gangnam style ai with my dog”
- “the Harry Potter Gangnam Style AI video, but with my photo”
Read more about where the trend came from on the Gangnam Style AI Video Generator page, or browse all meme video templates.