How to Make the APT AI Video With Two Photos

The APT AI trend takes a short clip styled after the music video for "APT." by ROSÉ and Bruno Mars and puts two new people in it. A bright pink studio, a drum kit and a wall of amplifiers; one performer sprawls on the floor singing into a microphone, the other drums behind, then both crowd the kit, belt the chorus and step out for the dance.

The version people are recasting is a cosplay take on that scene: a blue-haired singer in a witch hat and a red-haired drummer. APT AI replaces both of them with the people in your two photos and keeps everything else (every move, drum hit, cut, close-up and the sound) exactly where it is.

• Two photos, one person in each: the singer and the drummer. Not one group photo.
• A free account on this site, plus enough credits for one video.
• Any browser on a phone or computer. There is nothing to install, no prompt to write and no editing.

That's all. The scene, the choreography and the soundtrack are already loaded into the tool.

1. 1

   Open APT AI

   Go to the APT AI page and sign in. The original clip plays in the generator so you can see exactly what will be recreated.

2. 2

   Upload the singer's photo

   In the first box, add a photo of the person who should sing on the floor with the microphone (replacing the blue-haired singer).

3. 3

   Upload the drummer's photo

   In the second box, add a photo of the person who should play the drums (replacing the red-haired drummer).

4. 4

   Generate

   Click Generate. The video is 720p, 16:9 and costs 700 credits at 720p. You can close the tab while it renders.

5. 5

   Download and post

   When the video plays in the generator, download the MP4 and post it to TikTok, Instagram Reels or YouTube Shorts.

The order of the upload boxes decides the roles. Each box shows a still of the performer it replaces, so you can check before you generate.

Outfits come from your photos, not from the original. The witch hat and the cosplay costumes disappear, and each of you appears in what you wear in your own photo. If you want the costume look, wear it in the photo.

Because both of you are in almost every shot, both photos matter equally. A weak photo for either role shows up in most of the video.

• One person per photo, face clearly visible and facing the camera.
• Chest up or waist up gives the model the most detail; full body works too.
• Even light. Window light or daylight beats a dark room or a hard flash.
• No sunglasses, masks or heavy beauty filters. The model copies identity from the face.
• JPG, PNG or WebP, at least 341 px on each side and under 20 MB. A normal phone photo is fine.

A face-swap pastes your face over the original performer, so the hair, body and costume stay theirs. APT AI recasts the whole person, which is why the result reads as you rather than you wearing a mask.