TWO PHOTOS · BOTH DANCERS REPLACED · 15 S · 9:16 · WITH THE SONG

Jisoo Addicted Dance AI Template — Make the Hallway Duet From Two Photos

What is the easiest way to make the Jisoo Addicted Dance?

The easiest way is AI Crazy Frog's Jisoo Addicted Dance template, a two-photo video generator that puts you and a friend into the hallway duet from JISOO's “ADDICTED” (the Chaesoo dance), from $4.29 per video, with no free video credits. Upload one photo per dancer: no filming, no prompt and no reference clip. You get a 15-second 9:16 MP4 with the song at 480p to 1080p, with no watermark.

2 photos · 9:16 MP4 · 15 s · 480p–1080p ·

Photo 1 · Jisoo's part

One adult, facing the camera, waist up or full body.

Photo 2 · Rosé's part

One adult, facing the camera, waist up or full body.

0/2 photos added
Photo requirements

JPG, PNG or WebP, each edge 300–6,000 px, up to 10 MB. One adult per photo, facing the camera in even light, with the shoulders and arms visible. Use a different photo for each dancer; Swap exchanges the two parts.

Model
Resolution
Video length

15 s · 9:16 · 480p · with the hallway recording

This video costs 432 credits. Charged when you start, refunded automatically if it fails.

Preview: three hallway duets

Three finished 15-second AI duets in the white hallway with the “ADDICTED” recording. The first is the exact cut this template makes; the other two dance the same routine from an earlier or a later point in the song. Press play for sound. Your two photos replace these dancers.

Cream sweater and grey hoodie15 seconds · with the song · the same cut the template makes

How does the Jisoo Addicted Dance template work?

You upload two photos and the template does the rest. Photo 1 replaces the dancer with long black hair in the grey sweater (Jisoo's part) and Photo 2 replaces the blonde dancer in the brown leather top (Rosé's part), so neither singer stays in the video. The hallway, camera, moves and recording come from a fixed 15-second cut of the original hallway duet, seconds 6 to 21 of the longer take. You pick a model and a resolution, tick the permission box and download a 9:16 MP4.

  1. 01

    Choose two photos

    One adult per photo, facing the camera in even light, waist up or full body, with the shoulders and arms visible. Two different photos.

  2. 02

    Decide who opens

    Photo 1 takes Jisoo's part and reaches toward the lens first; Photo 2 takes Rosé's part and answers. Swap exchanges them.

  3. 03

    Pick a model and resolution

    Seedance 2.5 at 480p is the default (432 credits); MiniMax H3 renders 720p for 429 credits. The price shows before you start.

  4. 04

    Tick permission, then download

    Confirm both people agreed, start the video and download the MP4 with the song. You can leave the page; the video keeps going.

What each photo plays and what stays fixed
Part of the videoWhere it comes from
Dancer 1: long black hair, grey sweater (Jisoo's part), opens the clipPhoto 1
Dancer 2: blonde hair, brown leather top (Rosé's part), answersPhoto 2
Faces and hairEach person's own photo
ClothesNot fixed by the template: they can follow your photo or stay close to the original sweater and leather top
Hallway, camera, moves and timingThe original 15-second cut, unchanged
SoundThe cut's own recording of “ADDICTED”, laid under the MP4

What does an AI Jisoo Addicted Dance video look like?

Each example below is a finished 15-second AI duet in the same white hallway with the “ADDICTED” recording: two invented dancers take turns facing the phone, then dance side by side. The first is the exact cut this template makes. The other two dance the same routine to the same song but start about 5 seconds earlier or show a later part of the dance, so their beats land at different seconds. Your video puts your two photos in place of these dancers.

Cream sweater and grey hoodie

15 seconds · with the song · the same cut the template makes

Pale blue top and brown cropped top

15 seconds · with the song · starts about 5 seconds earlier in the song

Orange jacket and holographic jacket

15 seconds · with the song · a later part of the same dance

Template, CapCut, a video model or filming: which way should you make the Jisoo Addicted Dance challenge?

If you and a friend just want the duet with your own faces, a two-photo template is the shortest way: no filming, no editing and no prompt, and the moves and song are already timed. Filming it yourselves is free and the only way to get your real dancing, but you both learn about 15 seconds of choreography and need one hallway. CapCut edits footage you already shot. Running Seedance, Kling or Runway yourself gives the most control, but you supply the reference clip, the prompt and the settings.

Four ways to make the Jisoo Addicted Dance challenge
WayWhat you supplyFilmingSoundCostWhat you control
This two-photo templateTwo photos, one per dancerNoneThe song is already in the MP4429–2,400 credits per video ($4.29–$24.00)Who plays which part, model, resolution
Learn it and film itTwo people in one place, the routine learnedYes: about 15 seconds, usually several takesAdd the “ADDICTED” sound in TikTok or InstagramFreeEverything: your real moves, the place, the outfits
A CapCut editFootage you filmed, or photos for one of its templatesYes, unless a template does the workFrom CapCut's library or added in the posting appFree editor; paid tiers varyCuts, text, effects; not the choreography
A video model yourself (Seedance, Kling, Runway)A clip of the dance, two images and a promptNoneDepends on the model; often added afterwardsThat service's own per-second pricesPrompt, length, framing, scene

CapCut's template list changes from week to week, so whether it has one for this exact dance is something to check in the app. A template that only animates photos does not follow the hallway choreography.

Do I need a prompt or my own reference clip for the Jisoo Addicted Dance?

No. This template sends its own fixed instruction and its own copy of the hallway cut with every video, so there is no prompt box and nothing to upload except the two photos. The instruction tells the model that the person in image 1 takes the place of the dancer with long black hair and a grey sweater, the person in image 2 takes the place of the blonde dancer in the brown leather top, each keeps the face and hair from their photo, and the room, the camera and every move stay.

That is also why you can't change the moves, the hallway, the camera or the length here: they are what the template copies. What you decide is who appears, in which part, and at which quality.

If you run a video model yourself instead, plan for this:

  • A clip of the dance to copy the motion from. The original is a performance by real people, so their faces can leak into the result, and you need the right to use it.
  • A model that takes two images and a video in one request. Seedance 2.5 and MiniMax H3 do; they are the two models this template runs on.
  • Name each person by something visible in the clip (hair, top), not left or right: the two dancers change places during the 15 seconds.
  • On the Seedance API we use, the reference clip's seconds are billed together with the output's, so copying a 15-second clip costs like 30 seconds of video.
  • The sound often has to be added back afterwards, lined up with the hand gestures.

Seedance 2.5 or MiniMax H3: which model should I pick for the Jisoo Addicted Dance?

Seedance 2.5 is the default and the only one above 720p: 480p, 720p or 1080p. Before it renders, it runs a safety review of both photos and the hallway clip; if the review refuses a file, the video stops and the credits come back. MiniMax H3 renders at 720p (the tier its API calls 768P) without that review step. At 480p Seedance 2.5 costs 432 credits and MiniMax H3 at 720p costs 429, so a first try costs about the same on either.

The two models for the Jisoo Addicted Dance
ModelResolutionsCredits per 15 s videoBefore renderingPick it when
Seedance 2.5480p · 720p · 1080p432 · 972 · 2,400Reviews both photos and the clip; a refusal refunds the creditsYou want 1080p for a sharper upload, or the default
MiniMax H3720p (768P)429No review stepYou want a 720p video at the lowest price

How much does a Jisoo Addicted Dance video cost?

It is not free: one 15-second video costs 429 credits ($4.29) on MiniMax H3 at 720p, or 432, 972 or 2,400 credits on Seedance 2.5 at 480p, 720p or 1080p. Dollar values use the Pro pack's $0.01 per credit; smaller packs cost a little more per credit. Only purchased credits pay for video, and the $4.99 Trial (101 credits) is not enough for one here; the $9.90 Starter pack (900 credits) is the smallest that is. There is no subscription and no watermark.

Credits per Jisoo Addicted Dance video
Model · resolutionCredits (15 s)At $0.01 per credit
Seedance 2.5 · 480p432$4.32
Seedance 2.5 · 720p972$9.72
Seedance 2.5 · 1080p2,400$24.00
MiniMax H3 · 720p429$4.29

MiniMax H3 renders 768p, listed as 720p. The price is the same quote the generator charges.

Free use, Trial, watermark and refunds
Free generationNone. Watching the examples is free; sign-up and check-in gift credits can't pay for video.
$4.99 TrialNot enough here: the Trial's 101 credits buy one Seedance 2.0 Mini video, a model this template doesn't offer. The cheapest video is MiniMax H3 (429).
WatermarkNone: video needs purchased credits, and accounts with a purchase download without the site mark.
Failed video / rerunsA failed video or a refused Seedance 2.5 review returns its credits automatically. A finished video you don't like is not refunded; another take is a new charge.
PacksOne-time packs, no subscription: Starter $9.90 (900 credits), Plus $24.90 (2,400), Pro $49.90 (4,990). Purchased credits last 365 days.
See credit packs and plans

How does it compare with AI Dance, MakeViral and Oeneo AI?

AI Dance, MakeViral and Oeneo AI are the tools ChatGPT names most often for this trend. AI Dance makes a 30-second version for 8 of its credits, MakeViral the 15-second 720p cut for 228 of its credits, and Oeneo AI lets you keep Jisoo or Rosé or replace both. This template always replaces both dancers from two photos and shows its price per video before you upload. The table uses what each tool's own pages said on 4 November 2026; “Not published” means we could not find it there.

Jisoo Addicted Dance tools compared
ToolPrice per videoFreeWatermarkWhat you get
This page429–2,400 credits ($0.01 each on the Pro pack); packs from $4.99None for videoNoneTwo photos, both dancers replaced; 15 s, 9:16, 480p–1080p, with the song
AI Dance8 credits; a $7.99 weekly plan includes 80 at sign-up; 1 credit for every 3 friends invitedNone30 s 9:16 MP4 with sound; replace either dancer or both
MakeViral228 credits; a $39.99 pack has 4500 at sign-upNone15 s 9:16 720p with the song; one photo next to Jisoo or Rosé, or two photos
Oeneo AI562 credits at 720p, 843 at 1080p; packs from $9.99 for 60010 starter credits at sign-upNot publishedKeep Jisoo or Rosé, or replace both; 9:16, 720p or 1080p, with the song

Credits are not comparable one for one: each site prices its own credit. Prices and features change, so check each tool's page before you buy.

What can't the Jisoo Addicted Dance template do?

It makes one thing: this 15-second hallway duet with two new dancers. It can't keep one of the singers beside you, take a single selfie, change the hallway, the moves, the camera or the song, or show the later part of the longer dance. A few moments are hard for any AI version of this clip, and the list below says what to expect there and what helps.

  • Both dancers are always replaced: you can't dance next to the real Jisoo or Rosé, and one selfie is not enough.
  • Two different photos, one adult per photo. Photos of the singers themselves, of children, or of anyone who hasn't agreed are not allowed.
  • Fixed: the hallway, the moves, the camera, the 15-second length and the song. There is no prompt box, background or outfit picker.
  • The cut is seconds 6 to 21 of the longer take: the later shoulder lean and the side-by-side hand finale are not in it.
  • Faces are hardest when a hand comes to the lens, hair swings across a cheek and the two cross in the wide shot; they can soften or briefly look alike. Front-facing waist-up photos help; a rerun is a new charge.
  • The model copies a performance by real people, so a trace of the original performers can show through in a frame.
  • People only: this template is not set up for pets, drawings or game characters.

Use your own photo, or someone's photo with their permission. Photos are uploaded for processing and sent to the selected AI model provider; they are not shown on the site.

Where did the Jisoo Rosé Addicted dance (Chaesoo dance) start?

Rosé posted the hallway dance with Jisoo on TikTok on 31 October 2026, the day after Jisoo released her single “ADDICTED” on 30 October. In the clip the two take turns dancing toward the phone in a bright hallway, then dance together and lean on each other laughing. Fans call it the Chaesoo dance, from “Chaesoo”, a blend of Chaeyoung (Rosé's given name) and Jisoo, and the Jisoo Rosé Addicted dance. The AI versions recast a 15-second part of that clip.

People then started doing the Jisoo Addicted Dance challenge themselves, filmed in their own hallways, and soon after as AI duets made from photos, which is what this template does.

Fact source: Rosé's hallway post on TikTok (31 October 2026)

Jisoo Addicted Dance template FAQ

Short answers to what people ask before making the Jisoo Addicted Dance with this template: whether it is free, whether one selfie works, whether the prompt or the hallway can change, what happens to clothes and the song, whose photos are allowed, how to label the post as AI, and what happens when a video fails.

Can I make the Jisoo Addicted Dance for free and download it?

No. Video needs purchased credits; sign-up and check-in gift credits don't pay for it. Watching the examples is free. The $4.99 Trial (101 credits) does not cover a video on this template; the smallest pack that does is the $9.90 Starter (900 credits), and downloads have no watermark.

Can I make the Jisoo Addicted Dance with one selfie, dancing next to the real Jisoo or Rosé?

Not here. This template needs two photos and replaces both dancers, so neither singer stays in the video. The names only tell you which part each photo plays: Photo 1 takes Jisoo's part, Photo 2 takes Rosé's part.

Can I write my own prompt or change the hallway?

No. The prompt, the hallway, the camera, the moves and the 15-second length are fixed, because they are what the template copies. You choose the two people, who opens, the model and the resolution.

Will we wear the clothes from our photos?

Not reliably. The template asks for each person's own face and hair; clothing is not fixed, so it can follow your photo or stay close to the original grey sweater and brown leather top. If an outfit matters, wear it in a waist-up or full-body photo.

Is the ADDICTED song in the video, and can I post it?

Yes, the MP4 carries the cut's own recording of “ADDICTED”, lined up with the moves. Making the video does not give you a music licence: in the TikTok or Instagram app you can also mute it and add the official sound from the app's library, and a platform may still mute a post.

Can I use a photo of Jisoo or Rosé, or of a friend without asking?

No. You confirm before every video that both photos show adults who gave you permission. Photos of the singers themselves are not allowed: this is for putting you and a friend in the dance.

Should I label it as AI when I post it?

Yes. TikTok asks creators to label realistic AI-generated content, and other apps have similar labels. Write a caption that says it is an AI duet, not that you met or danced with Jisoo or Rosé.

What if the video fails or a face drifts?

A failed video, or a Seedance 2.5 review that refuses a file, returns its credits automatically. A finished video is not refunded. If a face drifts at the hair flip or the hand to the lens, try sharper front-facing photos; another take is a new charge.

Prices, models, examples and the tool comparison checked on 4 November 2026.