
Photo 1 — screams first, dances left
Opens the video in an extreme ultra-wide close-up: hands gripping the head, screaming into the lens, palms at the camera. From 7 s they dance in the background, then stand on the viewer's left in the courtyard.
Smalltown Boy AI video · two photos
Upload two photos, one person in each, to make the 24-second Smalltown Boy AI dance video. Photo 1 screams into the ultra-wide lens first (0–7 s) and dances on the left; Photo 2 takes the second close-up (7–14 s) and dances on the right, then both do the courtyard dance to Bronski Beat's "Smalltown Boy". Vertical 9:16, the song included, no prompt and no editing.
Make my Smalltown Boy AI video24 seconds · 9:16 · 480p · the song included
Each photo is first turned into a character sheet (front, back, profile, outfit), then the video follows the routine; on Seedance 2.5 the sheets, photos and clip are reviewed before the render. Usually a few minutes. The full 24 s is on Seedance 2.5; the 2.0 models make the first 5, 10 or 15 s.
This is the motion reference the model follows — both performers are bald, in black, so their looks can't leak into your video. Yours keeps the camera, cuts and moves with your two people, and the song.
The AI trend took off with cats: two TikTok remakes that keep the ultra-wide screams, the swap at 7 seconds and the courtyard dance. Play them with sound — this page makes the same routine with two people.
These videos were made by TikTok creators with other tools and are shown, with credit and a link to each post, as examples of the trend — they were not made on DanseAI, and this page does not make pet versions.
A Smalltown Boy AI video recasts the viral scream-swap-dance clip with two people from your photos: each screams into an ultra-wide lens in turn, then both dance to Bronski Beat's "Smalltown Boy". On this page it takes two photos, runs 24 seconds (or the first 5, 10 or 15) in 9:16, includes the song and costs 10 to 747 credits depending on the model, length and resolution.
Last verified: October 10, 2026
Yes. The model renders a silent video, then the Bronski Beat "Smalltown Boy" track from the reference is laid underneath, so the screams, the swap and the dance land on the music in the MP4 you download.
The recording belongs to its rights holders; including it in your file is not a licence to redistribute it. TikTok or Instagram may match the track when you post — you can also mute it and pick the song from the app's sound library.
It started as a real video: on 4 August 2026 Elvira (@elvira.mua.music) and dancer @mr_unvrs posted a clip on Instagram in which each screams into a wide-angle phone held inches from the face, they swap, then dance side by side to "Smalltown Boy".
The AI wave came on 2 October 2026, when @bellvtrix.ai posted a version with two cats made with Higgsfield Genjutsu; it passed a million likes on TikTok. Since then people recast the dance with friends, partners, parents and characters, using AI video tools that keep the original camera and cuts and swap the people.
"Smalltown Boy" is the 1984 debut single of the British synth-pop trio Bronski Beat, from the album The Age of Consent, sung in Jimmy Somerville's falsetto. It reached the UK top three; the song tells of a young gay man leaving a hometown that doesn't accept him.
"Smalltown Boy" also backed a 2024 TikTok trend, "Dance like it's the 80s", where parents showed off their old moves, and Just Dance has its own routine for the song. This page makes neither: it makes the 2026 scream, swap and dance duo.
Sources:
DanseAI is an independent tool. It is not affiliated with, endorsed by or sponsored by Bronski Beat, its members or labels, the original video's creators, eachlabs.ai or Higgsfield. The template is a motion reference plus its audio, made for parody and personal posts; use photos of yourself or of people who agreed to appear.
No prompt, no editing and no clip to find: the routine is built in. It usually takes a few minutes and you can leave the page while it renders.
One person per photo, face sharp and forward, waist-up or full-body if you can. Photo 1 screams first and dances on the left; Photo 2 screams second and dances on the right. Press Swap to flip them.
Keep the original courtyard or choose another place. 24 s is the whole routine on Seedance 2.5; 5, 10 or 15 s works on every model. The exact credit cost shows next to the Generate button.
Play it with sound, check each face in its close-up and in the dance, then download the 9:16 MP4 and post it to TikTok, Reels or Shorts.
The two roles are fixed by the order of your photos — not by gender, height or age. Two friends, a couple, siblings or you and a parent all work.

Opens the video in an extreme ultra-wide close-up: hands gripping the head, screaming into the lens, palms at the camera. From 7 s they dance in the background, then stand on the viewer's left in the courtyard.

Dances behind Photo 1 during the first scream, then takes the second close-up at 7 s: hands on the head, shocked, sliding down to the cheeks. Stands on the viewer's right for the dance and the final lunge.
Six moments in the full 24 seconds, with the reference's own cuts. Shorter videos are the opening: 5 s is Photo 1's scream, 10 s adds the swap, 15 s stops as the dance begins.
An extreme ultra-wide close-up: both hands on the head, mouth wide open, while Photo 2 dances in the background.
Both open hands shoved at the lens, then a wide-eyed lean in.
Hard cut to Photo 2's close-up: hands on the head, screaming in shock, sliding down to the cheeks. Photo 1 now dances behind.
A wide shot from slightly above: the two face each other and swing and cross their arms in a quick hand-jive, Photo 1 on the left.
The camera drops to the ground and looks up as both bounce in a half-crouch, throwing their arms up against the sky.
Both stand wide-eyed with mouths open, then rush the lens together, hands reaching for the camera.
Close-ups through an ultra-wide lens distort faces, and the two people trade places twice. The generator splits the job into four inputs, each with one job.
Before the video, each photo becomes a sheet: front and back full body, a portrait and three head angles in the same outfit. It gives the model your back, profile and shoes, which a single selfie doesn't show.
Sent alongside the sheets as the authority for each face, so eye colour, facial hair, skin tone and hair stay yours — a clean-shaven face stays clean-shaven.
The camera, cuts and choreography come from a 24 s re-render of the original dance in which both performers are bald and dressed in black, so neither of their faces, hair nor outfits can leak into yours.
Keep the slatted-fence courtyard, or send a scene image after the photos: the moves stay, the place changes.
Original courtyard
Red fence garden
Living roomOr upload a photo of your own empty place — a garden, a hallway, a stage. Any people in it are left out of the video.
Credits per video by model, length and resolution, including both character sheets (live prices, the same for every account):
| Model · length | 480p | 720p | 1080p |
|---|---|---|---|
| Seedance 2.0 Mini · 5s | Costs 10 credits | Costs 15 credits | — |
| Seedance 2.0 Mini · 10s | Costs 14 credits | Costs 24 credits | — |
| Seedance 2.0 Mini · 15s | Costs 18 credits | Costs 33 credits | — |
| Seedance 2.0 Fast · 5s | Costs 21 credits | Costs 40 credits | — |
| Seedance 2.0 Fast · 10s | Costs 37 credits | Costs 74 credits | — |
| Seedance 2.0 Fast · 15s | Costs 53 credits | Costs 109 credits | — |
| Seedance 2.0 · 5s | Costs 32 credits | Costs 63 credits | Costs 150 credits |
| Seedance 2.0 · 10s | Costs 59 credits | Costs 121 credits | Costs 295 credits |
| Seedance 2.0 · 15s | Costs 86 credits | Costs 178 credits | Costs 440 credits |
| Seedance 2.5 · 5s | Costs 44 credits | Costs 92 credits | Costs 160 credits |
| Seedance 2.5 · 10s | Costs 83 credits | Costs 179 credits | Costs 314 credits |
| Seedance 2.5 · 15s | Costs 121 credits | Costs 267 credits | Costs 469 credits |
| Seedance 2.5 · 24s | Costs 191 credits | Costs 424 credits | Costs 747 credits |
A Smalltown Boy AI video costs 10 to 747 credits, depending on the model, length and resolution. The exact price is shown before you press Generate. The full 24-second routine at 480p is 191 credits; one-time credit packs start at $12 for 220 credits, with no subscription.
For scale: the full routine at 480p is 191 credits, about $7.10 of credits on a monthly plan — a credit value, not a separate per-video price.
Every run comes out a little different. If a face drifts, try sharper photos or Seedance 2.5 — failed generations cost nothing.
There are a few ways to make the Smalltown Boy AI dance. As of October 10, 2026, this is how they differ:
| Tool | You upload | Video | The routine | Song |
|---|---|---|---|---|
| DanseAI (this page) | Two photos; a character sheet is made from each | 24 s or the first 5/10/15 s; 9:16; 480p–1080p; four Seedance models; four background choices | Built in: both screams, the swap, the courtyard dance and the lunge | Included |
| Other Smalltown Boy template sites | Two photos (some take pets) | Usually one fixed length of about 21–25 s, in 9:16 or 3:4 | Built in | Some include it, some export silent |
| Motion-transfer tools (e.g. Higgsfield Genjutsu) | The clip you find and trim plus your images | Follows the clip you supply | Copied from your clip; the original dancers' looks can leak through | From your clip or added yourself |
| A video model with a text prompt | Photos and a long prompt | Your settings | The swap and the exact cuts are hard to get from text | Add the song yourself |
If you just want the trend with two friends in a few minutes, a two-photo template is the shortest path; motion-transfer tools suit other clips or pets.
An exact likeness isn't guaranteed: the ultra-wide lens stretches faces, and a face can drift for a moment in the fast cuts. These habits keep the video closest to your photos.
The same photo twice is accepted, but the template is made for two different people.
The trend lives on its caption. Tap one to copy it with #SmalltownBoy #BronskiBeat.
Every credit pack and plan works on this page and on the Alors on Danse, Hotel Lobby, Rumpelstiltskin and Burning Bridges AI generators. Any purchase removes the watermark.
Starter
For trying the trend
240 credits / month
≈ 10 ten-second Seedance 2.0 Mini videos
Pro
PopularFor regular creators
780 credits / month
≈ 32 ten-second Seedance 2.0 Mini videos
Enterprise
For agencies and pages
2,650 credits / month
≈ 110 ten-second Seedance 2.0 Mini videos
This page does exactly that. Upload two photos, one person each, pick the background, length and model, and press Generate: the video follows the original scream, swap and courtyard dance with your two faces, and the Bronski Beat song is included.
Put your photo in one slot and your friend's in the other, one person per photo, with their permission. Photo 1 screams first and dances on the left, Photo 2 second and on the right; the character sheets are made from the photos automatically, so you don't upload extra angles.
Yes — this website. It runs in the browser on a phone or computer, so there is nothing to install. Upload two photos, generate, and download a 9:16 MP4 with the song for TikTok, Reels or Shorts.
The viral two-cat version that started the AI trend was made with Higgsfield Genjutsu, a motion-transfer tool. Many people now use dedicated two-photo template pages that keep the original camera and cuts and swap the people. This page is one of them and runs on Seedance in reference mode.
Yes. There is no clip to find, no prompt to write and no timeline to edit: the screams, the swap at 7 seconds, the dance and the song are built into the template. You upload two photos and download the finished video.
Faces hold up best when the model sees more than one selfie, so this page first turns each photo into a character sheet (front, back, profile, head angles, outfit) and sends the sheets with the photos. The ultra-wide close-ups still stretch faces and a fast cut can make one drift for a moment, so it isn't perfect every time; sharp, well-lit photos and Seedance 2.5 come closest.
Not entirely free here: creating an account and watching the examples are free, and each video costs credits, shown before you generate. Signup credits and the $5 trial can pay for a Seedance 2.0 Mini video. Videos made before your first purchase carry a small moving watermark; any purchase unlocks the clean original, including videos you already made.
Not on this page: the template is made for two people and the character sheets assume a person. The cat versions on TikTok were made with motion-transfer tools such as Higgsfield Genjutsu.
None here. The routine, camera, cuts and timing are built in. A text prompt alone usually gives one person screaming at a fisheye camera and misses the two-person swap at 7 seconds and the exact cuts, which is why the template follows a motion reference instead.
All of them appear: the viral cat version came from Higgsfield Genjutsu, and other generators use their own models. This page uses Seedance in reference mode — Seedance 2.5 for the full 24 seconds and the best likeness, Seedance 2.0, 2.0 Fast or 2.0 Mini for the shorter openings.
Yes. Upload the same photo in both slots and that person screams twice and dances with themselves. The template is made for two different people, so a solo version is a fun variation rather than the trend's usual look.
Photo 1 screams first (0–7 s) and stands on the viewer's left in the dance; Photo 2 takes the second close-up (7–14 s) and stands on the right. Roles follow photo order, not gender, so any two people work. Press Swap to flip them.
Yes. The "Smalltown Boy" track from the reference is laid under your video in time with the moves. The music belongs to its rights holders, so follow each platform's music rules when you post.
A Smalltown Boy AI video costs 10 to 747 credits, depending on the model, length and resolution. The exact price is shown before you press Generate. The full 24-second routine at 480p is 191 credits; one-time credit packs start at $12 for 220 credits, with no subscription.
24 seconds is the whole routine and is made on Seedance 2.5; every model can make the first 5, 10 or 15 seconds. All videos are vertical 9:16. 1080p is available on Seedance 2.5 and 2.0; Fast and Mini stop at 720p.
Yes. Keep the original courtyard, pick the red fence garden or the living room, or upload a photo of your own empty place. The moves and camera stay the same; only the setting changes.
From a real Instagram video posted on 4 August 2026 by Elvira (@elvira.mua.music) and @mr_unvrs. It went viral as an AI trend after @bellvtrix.ai posted a two-cat version made with Higgsfield Genjutsu on 2 October 2026.
No. If a character sheet, the review or the render fails, the credits come back automatically. Try a clearer photo of one real person.
Two photos, the courtyard and the song. Pick 24 seconds or a shorter opening, see the price, press Generate and post your Smalltown Boy AI video in minutes.
Make my Smalltown Boy AI video