Skip to content

ROSÉ and Bruno Mars, the pink studio

APT AI Video From Two Photos

Put yourself and a friend into the APT. music video: photo 1 plays the drums, photo 2 sings, and the MP4 comes with the song. No filming, no prompt, no app.

  • Photo 1 drums, photo 2 sings
  • 15 seconds with the song
  • Shot for shot, 16:9
  • 6 credits at 480p
1

Your photo

One photo is enough: solo or the two of you.

2

Effect

APT. is made for two people: add a second photo, or use one photo of you both.

3

Video

This video: 6 credits. Ready in about 2 to 5 minutes; failed generations return your credits. View pricing

The short answer

Hotel Lobby Video (hotellobbyvideo.net) makes the APT AI video from two photos: the person in photo 1 plays the drums and the person in photo 2 sings in the pink studio of ROSÉ and Bruno Mars’s APT. music video, 15 seconds shot for shot in 16:9, and the downloaded MP4 comes with the matching 15 seconds of the song. It runs on Seedance 2.0 Mini: 6 credits at 480p or 12 at 720p, packs start at 10 credits for $9.99, there is no free preview, and a failed render is refunded. The provider quotes up to about 5 minutes.

Updated

The trend

What an APT AI video is

“APT.” is the duet ROSÉ and Bruno Mars released on 18 October 2024, named after a Korean drinking game (“apateu” is Korean for apartment). In the official video the two share a pink studio: one at the drums, one singing, trading close-ups and dancing side by side. The AI version recasts that scene with two other people.

Here it is a template in the generator: APT. under Effects. You add a photo of each person, and the video follows the official one from 0:32 to 0:47, shot by shot: kneeling by the drum kit, two narrow camcorder shots with black bars, the drums, a close-up of a singing mouth and a wide shot of the two of you dancing in front of the amplifiers. Faces, hair and clothes come from your photos.

Three steps

How to make the APT. video with a friend

  1. Frame from our APT. demo: the singer kneels by the red drum kit and sings into a microphone.
    Frames from our demo
    01

    Add two photos

    Photo 1 is the drummer, photo 2 the singer: one clear, front-facing photo per person. One photo of both works too (left drums, right sings).

  2. Frame from our APT. demo: a narrow camcorder shot with black bars, the two performers close together.
    Frames from our demo
    02

    Keep APT. selected

    The template is always 15 seconds and 16:9. It starts on Seedance 2.0 Mini at 480p; pick any Seedance model or resolution and the price on the button follows.

  3. Frame from our APT. demo: both performers dance side by side in front of black amplifiers.
    Frames from our demo
    03

    Create and download

    The provider quotes up to about 5 minutes; our tests took about 3 minutes. The MP4 already has the song in it, ready to post.

What you get

What is in your video

Frame from our APT. demo: the pink studio from above, with the lightning-bolt drum kit and the amplifiers.
Frames from our demo
01

The pink studio, shot for shot

The shots and cuts of the official video: the red drum kit with its lightning bolt, the black amplifiers, the pink floor.

Frame from our APT. demo: a close-up of the singer, with the face from the photo.
Frames from our demo
02

Your faces, not theirs

The model follows a line drawing of the clip with every face erased, so the faces come from your photos alone.

Frame from our APT. demo: the extreme close-up of a singing mouth.
Frames from our demo
03

The song in the MP4

The matching 15 seconds of the song, laid under the video after it is rendered.

Frame from our APT. demo: the second camcorder shot, a hug at the drums with black bars.
Frames from our demo
04

Its camcorder moments

The two narrow handheld shots keep their black side bars, as in the official video.

Our test render

This is our second test, made on 12 October 2026 from two AI-made people with the same prompt and reference clip the generator sends: Seedance 2.0 Mini, 480p, 15 seconds, 16:9. Both faces stayed theirs from the first frame to the last, with none of the artists’ looks, outfits or sunglasses, and the set and the order of the shots follow the official video. Our first test put the black bars on the wrong shots and added a cap in the last close-up; the wording the generator uses now fixed both. In the mouth close-up the edge of a microphone still shows.

Our render

Seedance 2.0 Mini · 480p · 15 s · 16:9

  • Product test render, made outside the website with our generation pipeline.
  • Audio: the trend's own song, cut to the clip and added after the silent render.
  • Web preview: 864 × 496, compressed for playback. The settings above describe the original render.

What happens in the 15 seconds

TimeShot
0 to 3.3 sHigh angle: the singer kneels by the drum kit with a microphone, the drummer plays behind it
3.3 to 6.7 sNarrow camcorder shot with black bars: a close two-shot, then the drums
6.7 to 9.9 sMedium shot: the drummer plays, the singer dances and hugs them
9.9 to 11.4 sNarrow camcorder shot with black bars: a hug at the drums
11.4 to 13 sExtreme close-up of the singer’s mouth
13 to 14.6 sWide shot: both dance in front of the amplifiers
14.6 to 15 sClose-up of the singer

How the template makes the AI video

Tools that make this trend work in one of three ways: an image-to-video model with a long shot-by-shot prompt, a template that renders the scene again with your faces, or a face swap on the original footage. This is a template. The model never gets the official video: it gets a line drawing of the clip with every face erased, the song for timing, and your photos. It renders without sound, and the song is laid under the finished video, so the music is the real recording, not a re-sung copy.

Tips and limits

  • Use sharp, front-facing photos without sunglasses; waist up or full length works best.
  • It is 16:9 only. For a vertical post, crop it in TikTok or CapCut; there is no 9:16 version.
  • One photo of both is supported, but our tests used two photos.
  • Pets have not been tested in this template.
  • Only use photos of people who agreed: no celebrities, no strangers.
  • The song belongs to its rights holders. An app may mute or flag a post; you can also post the video and pick the APT. sound inside the app.

Made here

Hotel Lobby Video Examples

Two selected Wan 3.0 results from our generation workflow. Each rap duo video lists its settings and whether it was a product test or a website render. Play each clip with sound and review the whole performance; these examples do not establish a success rate.

Our render

Wan 3.0 · 480p · 10 s · 16:9

  • Product test render, made outside the website with our generation pipeline.
  • Audio: Hotel Lobby template track.
  • Web preview: 854 × 480, compressed for playback. The settings above describe the original render.
Our render

Wan 3.0 · 720p · 12 s · 16:9

  • Generated through this website.
  • Audio: Hotel Lobby template track.
  • Web preview: 854 × 480, compressed for playback. The settings above describe the original render.

From other creators

Trend Examples from Other Creators

These clips illustrate the Hotel Lobby format and were made with other tools. They are references, not evidence of this generator’s output. Each links to its credited creator.

Questions

APT AI video: questions

Still stuck? Contact us or see pricing.

What is the APT. trend?

People put themselves and a friend into the pink-studio scene of ROSÉ and Bruno Mars’s APT. music video, one on the drums and one singing. Here it is made from two photos: 15 seconds, shot for shot, with the song.

How do I make the APT. video with my friend or partner?

Add a photo of each of you (photo 1 drums, photo 2 sings), keep APT. selected and tap create. A few minutes later you download an MP4 with the song in it.

Does it come with the song?

Yes. The downloaded MP4 has the matching 15 seconds of the song. The video is rendered without sound and the real recording is added afterwards, so nothing is re-sung.

Will we turn into ROSÉ and Bruno Mars?

Not in our tests. The model sees a drawing of the clip with the faces erased, so faces, hair and clothes come from your photos. Results vary from render to render: a failed render is refunded, a finished one you do not like is not.

Who plays the drums and who sings?

Photo 1 plays the drums and photo 2 sings; swap the photos to swap the parts. With one photo of both, the person on the left drums and the person on the right sings.

Can I make it vertical for TikTok or Reels?

Not here: the template is 16:9, like the official video. Crop it in TikTok, Instagram or CapCut if you want it vertical.

How much does it cost? Is it free?

It is not free and there is no free preview. 15 seconds cost 6 credits at 480p on Seedance 2.0 Mini or 12 at 720p; other Seedance models cost more, and the button shows the price. Credits are paid when you create; the smallest pack is 10 credits for $9.99, and a failed render is refunded automatically.

Do I need an app or a prompt?

No. It runs in the browser on a phone or a computer, and the scene, the moves and the song are part of the template. You only add the photos.

Can I use a photo of a celebrity?

No. Only use photos of yourself and of people who agreed. Photos of celebrities, public figures or strangers are not allowed.

Is this affiliated with ROSÉ or Bruno Mars?

No. Hotel Lobby Video is an independent tool and is not affiliated with ROSÉ, Bruno Mars or their labels.

Grab the drumsticks

Make your video

Two photos, the pink studio and the song. The price is shown before you create.