Smalltown Boy AI video generator
Smalltown Boy AI Video Generator: Put Your Duo in the Scream-and-Dance Trend
Upload two photos, one person in each, to make your own Smalltown Boy AI video without writing a prompt. In the full 24-second version, Photo 1 takes the first scream (0–7 s) and dances on the left, Photo 2 takes the second scream (7–14 s) and dances on the right, then both lunge at the lens to Bronski Beat's “Smalltown Boy”.
Make my Smalltown Boy AI video- Two photos · no prompt
- 5, 10, 15 s or the full 24 s · 9:16
- Smalltown Boy soundtrack added
- People only · credits shown first
24 s · 9:16 · 480p · Smalltown Boy soundtrack added after generation
Each photo first becomes a character sheet (included in the price), then the video is made. Seedance 2.5 holds faces best and is the only model for the full 24 s; Mini is the cheapest way to test a pair. Credits are charged when the job starts and refunded automatically if it fails.
Reference clip: the 24-second scream-and-dance routine with both performers re-rendered bald and in black, so their looks never carry over. Your faces come from the photos, hair and outfits from the character sheets made from them.
Smalltown Boy AI Video Generator: Quick Facts
Everything the generator does, in one table. Checked against this page's own generator on 10 October 2026.
| Item | Smalltown Boy AI video on Raindance Maker |
|---|---|
| Input | 2 photos, one person in each (JPG, PNG or WebP) |
| Photo 1 | First scream close-up (0–7 s), then dances on the viewer's left |
| Photo 2 | Second scream close-up (7–14 s), then dances on the right |
| Roles | Set by upload order, not by gender; swap the photos to swap the roles |
| How it is made | Each photo becomes a character sheet (GPT Image 2.5), then Seedance reference-to-video or MiniMax H3 follows a fixed 24-second motion reference |
| Prompt | Not needed; the routine and camera come from the reference |
| Lengths | 5, 10 or 15 s (the opening of the routine) or the full 24 s |
| Full 24-second routine | Seedance 2.5 only |
| Models | Seedance 2.5 (default), Seedance 2.0 Mini, 2.0 Fast, 2.0 and MiniMax H3 |
| Format | 9:16 vertical MP4: 480 × 854 (default), 720 × 1280 or 1080 × 1920 |
| Resolution limits | Mini and Fast up to 720p; MiniMax H3 720p or 1080p |
| Background | Original courtyard, red fence garden, living room or a picture of your own |
| Soundtrack | Rendered silent, then the reference clip's “Smalltown Boy” audio is added |
| Cheapest video | 8 credits (Seedance 2.0 Mini, 5 s, 480p) |
| Full 24 s on Seedance 2.5 | 190 credits at 480p, 422 at 720p, 746 at 1080p |
| Watermark | Free accounts get a moving raindancemaker.com watermark; any paid order unlocks the clean original |
| Failed or rejected jobs | Credits refunded automatically |
| Cats and dogs | Not supported; people only |
| Who sees the result | Only you, in your account |
The exact credit cost always shows on the Generate button before you start.
How to Make a Smalltown Boy AI Video from Two Photos
1. Upload two photos
Put the person who should scream first in Photo 1 and the second screamer in Photo 2. One real person per photo, face visible. Use the swap button to change who goes first.
2. Pick a length, model and background
Choose 24 s on Seedance 2.5 for the whole routine, or 5, 10 or 15 s on any model for the opening. Keep the original courtyard or pick another scene, then check the credit cost on the button.
3. Generate and download
Each photo becomes a character sheet, the video is made without sound and the Smalltown Boy audio is laid back on. Check both people, then download the vertical MP4 for TikTok, Reels or Shorts.
You don't upload a reference video or write a prompt: the routine, the cuts and the camera are already set up.
Who Screams First? Photo 1 vs Photo 2
The Smalltown Boy routine has two people, and each keeps the same role from the first scream to the final lunge. Here is what each photo does in the full 24-second video.
| Time | What happens | Who |
|---|---|---|
| 0:00–0:07 | First ultra-wide close-up: hands gripping the head, screaming, palms shoved at the lens | Photo 1 (Photo 2 dances small in the background, left) |
| 0:07–0:14 | Second close-up: screaming in shock, hands sliding down to the cheeks | Photo 2 (Photo 1 dances behind, right) |
| 0:14–0:24 | Hand-jive in the courtyard, low-angle bounce, a frozen stare, then both lunge at the camera | Photo 1 on the left, Photo 2 on the right |
Roles follow the upload order, not gender: two guys, two girls, a couple or a parent and child all work. Swapping the photos swaps who screams first and which side each person dances on.
The reference is the public 24-second clip of eachlabs.ai's Smalltown Boy dance flow, re-rendered from the original dance by Elvira and mr_unvrs on Instagram
The Smalltown Boy Dance Moves, Shot by Shot
Every Smalltown Boy AI video follows the reference cut for cut. These are the four moments the trend is known for.
0:01The first scream
Photo 1 grips their head and screams into the ultra-wide lens, so close the face fills the frame.
0:09The swap
The camera turns to Photo 2 for the second scream while Photo 1 keeps dancing in the background.
0:15The courtyard dance
Side by side by the ivy fence, both swing and cross their arms in a quick hand-jive, then bounce in a low crouch.
0:21The lunge
After a frozen, wide-eyed stare, both rush the lens with their hands reaching out, faces filling the frame.
Stays from the reference
The moves and their timing, every cut, the handheld ultra-wide camera and its distortion, the vertical framing and, unless you pick another scene, the courtyard.
Comes from your photos
Faces, skin tone, hair (or a bald head, if that is what the photo shows), facial hair, glasses and clothes. Nothing about how the reference performers look is copied.
Smalltown Boy AI Video Lengths: 5, 10, 15 or 24 Seconds
Short versions are the opening of the reference, not the whole routine squeezed into fewer seconds. Choose 24 seconds if you want the duo dance and the ending.
| Length | What it shows | Models |
|---|---|---|
| 5s | Photo 1's scream close-up only; Photo 2 appears small, dancing in the background | All five models |
| 10s | Photo 1's close-up and the first three seconds of Photo 2's | All five models |
| 15s | Both scream close-ups and the first second of the courtyard dance | All five models |
| 24s | The whole routine: both screams, the dance, the low angle, the freeze and the lunge | Seedance 2.5 only |
What AI Model Makes the Smalltown Boy Video?
No single official model sits behind every Smalltown Boy video. Versions shared on TikTok were made with motion-transfer and reference-video tools, and the names that come up most are Higgsfield's Genjutsu, Kling Motion Control and Seedance. A plain text-to-video prompt cannot copy the routine, and CapCut effects that only paste the original dancers over a background don't put your faces in it.
This generator works in two steps. First, GPT Image 2.5 turns each photo into a character sheet: front and back full body, a large portrait and three head angles, in the photo's own clothes. Then Seedance reference-to-video (or MiniMax H3) follows the 24-second motion reference, taking the faces from your photos and the bodies and outfits from the sheets.
Why the sheets: the reference performers are bald and dressed in black so their looks don't leak into your video, but that also means the model has nothing to copy your hair, back view or shoes from. The sheet fills that gap. It is included in the price and you never have to look at it.
Which model to pick
- Seedance 2.5: best likeness in the extreme close-ups and the only model for the full 24 s; 480p, 720p or 1080p.
- Seedance 2.0 Mini: the cheapest way to test a pair of photos, up to 15 s at 480p or 720p.
- Seedance 2.0 Fast: a middle option, up to 15 s at 480p or 720p.
- Seedance 2.0: up to 15 s, also at 1080p.
- MiniMax H3: up to 15 s at 720p or 1080p.
Results vary from run to run. The models are built to keep each identity, but a face can still drift in the most stretched close-up frames.
Smalltown Boy AI Video Pricing in Credits
The price depends on the length, the model and the resolution, and already includes both character sheets. The generator shows the same number before you start.
| Model · resolution | 5s | 10s | 15s | 24s |
|---|---|---|---|---|
| Seedance 2.5 · 480p | 43 | 81 | 120 | 190 |
| Seedance 2.5 · 720p | 91 | 178 | 265 | 422 |
| Seedance 2.5 · 1080p | 158 | 313 | 468 | 746 |
| Seedance 2.0 Mini · 480p | 8 | 12 | 17 | — |
| Seedance 2.0 Mini · 720p | 13 | 22 | 32 | — |
| Seedance 2.0 Fast · 480p | 20 | 36 | 52 | — |
| Seedance 2.0 Fast · 720p | 38 | 73 | 107 | — |
| Seedance 2.0 · 480p | 31 | 58 | 85 | — |
| Seedance 2.0 · 720p | 62 | 119 | 177 | — |
| Seedance 2.0 · 1080p | 149 | 294 | 439 | — |
| MiniMax H3 · 720p | 42 | 81 | 119 | — |
| MiniMax H3 · 1080p | 65 | 127 | 188 | — |
At the cheapest monthly-plan credit rate ($0.037), the full 24 s on Seedance 2.5 works out to about $7.06 at 480p, $15.69 at 720p and $27.74 at 1080p. That is a credit rate, not a single-video price: credits come in packs and plans.
The $5 trial pack (24 credits) covers Seedance 2.0 Mini at 480p up to 15 s, but not 24 s. The Starter plan (240 credits for $9 a month) covers one full 24-second video at 480p.
What uses credits, and what doesn't
- The two character sheets cost nothing extra; they are part of each quote.
- New accounts may receive 5 signup credits to try it.
- Videos made on a free account carry a moving raindancemaker.com watermark.
- Any paid order removes the watermark from new videos and unlocks the clean originals of the ones you have already made.
- Credits are charged when a job starts. If it fails, they are refunded automatically.
Compare plans on the pricing page
Tip: test a new pair on Seedance 2.0 Mini at 5 s first, then make the 24-second Seedance 2.5 video once you like the casting.
Credits for Smalltown Boy AI videos
Every plan works for Smalltown Boy AI, the Raindance generator, Hotel Lobby AI, Alors on Danse AI and Rumpelstiltskin AI. Any paid order removes the watermark from new videos and unlocks the clean originals of earlier ones.
Starter
For trying the trend
240 credits / month
≈ 2 ten-second MiniMax H3 720p videos
- Watermark-free videos, including ones you already made
- Failed videos refunded
- Email support
Pro
For regular creators
780 credits / month
≈ 9 ten-second MiniMax H3 720p videos
- Watermark-free videos, including ones you already made
- Failed videos refunded
- All lengths up to 15 seconds
- Priority support
Enterprise
For agencies and pages
2,650 credits / month
≈ 32 ten-second MiniMax H3 720p videos
- Everything in Pro
- Dedicated support
- Custom integrations
Is There a Free Smalltown Boy AI Video Generator Without a Watermark?
Not here, and be wary of claims elsewhere. A Smalltown Boy AI video is a paid video-model job, so every one costs credits.
Videos made on a free account carry a moving raindancemaker.com watermark. After any paid order, new videos are clean, and the clean original of a video you already made unlocks without generating it again.
New accounts may get 5 signup credits, enough to try a short Seedance 2.0 Mini clip but not the full 24 seconds.
If a job fails or the provider rejects a photo, the credits come back automatically.
Does the Smalltown Boy AI Video Include the Original Song?
Yes. The video model renders without sound, then the same seconds of the reference clip's audio, Bronski Beat's “Smalltown Boy”, are mixed back in, so the music lands on the same moves. You don't need to add the sound in TikTok.
The song and recording belong to their rights holders. Having the audio in your file is not a licence to distribute or monetise it, and TikTok, Instagram or YouTube may mute or restrict it. If that happens, use the platform's own music library.
What the Smalltown Boy AI Generator Can and Cannot Do
| Feature | Supported |
|---|---|
| Two people from your own photos | Yes |
| The full 24-second routine | Yes, on Seedance 2.5 |
| Vertical 9:16 for TikTok and Reels | Yes |
| Smalltown Boy audio in the file | Yes, subject to music rights and platform rules |
| Another background | Yes: two preset scenes or your own picture |
| Cats or dogs | No |
| Three or more people | No |
| Another song | No |
| Different moves or a custom prompt | No |
| Results visible only to you | Yes |
Best Photos for the Smalltown Boy AI Dance
The routine opens on extreme close-ups, so the photos decide how convincing the faces look.
- One person per photo. Crop out anyone else; a second face can end up in the courtyard.
- Face the camera. A clear, front-on face without sunglasses or a hand in the way holds up best in the scream.
- Show the outfit. A half- or full-body photo gives the character sheet the real clothes and shoes; a head-only photo gets a plain outfit in matching colours.
- Even light, no heavy filters. Strong filters and harsh shadows tend to show up in the video.
- Ask first. Use yourself or adults who agreed to be in the video, not celebrities or strangers.
- Run it again if needed. Each generation varies a little; the same pair can come out better on a second run or another model.
Duo ideas
- You and your best friend
- Siblings, the loudest one first
- A couple, whoever screams better in Photo 1
- Two co-workers before the weekend
- Roommates on move-out day
What Is the Smalltown Boy AI Video Trend?
This page is an independent tool. It is not affiliated with the original dancers, Bronski Beat, eachlabs.ai or TikTok.
The Smalltown Boy AI trend recasts one short dance video: two people take turns screaming into an ultra-wide phone lens, then dance together in a courtyard and rush the camera, set to Bronski Beat's “Smalltown Boy”. People remake it with themselves, friends, celebrities and pets.
It started as a real dance clip by Elvira (@elvira.mua.music) and mr_unvrs, posted on Instagram on 4 August 2026. AI remakes followed on TikTok, including widely shared two-cat versions, and the Smalltown Boy dance trend took off from there.
“Smalltown Boy” is Bronski Beat's 1984 synth-pop debut single, sung by Jimmy Somerville in his falsetto. It tells of a young gay man who leaves his hometown after his family rejects him. The trend uses the instrumental build and the chorus.
Not to be confused with
Older TikTok videos of parents dancing to the song, the original 1984 music video, or a Just Dance routine. This page makes only the scream-and-dance duo video.
Sources: the original dance on Instagram · the eachlabs.ai reference flow
Smalltown Boy AI Video FAQ
What is the best Smalltown Boy AI video generator right now?
It depends on what you need. For a two-person version from your own photos with the full 24-second routine, this Raindance Maker generator uses Seedance 2.5 and adds the song; shorter 5, 10 or 15-second versions run on cheaper models. It does not make pet versions.
Where can I create the Smalltown Boy AI video with my own photos?
Right here: upload two separate photos, one real person in each, pick a length and model, and press Generate. The routine and camera come from the built-in reference, so there is no prompt and no video to upload.
Is there a free Smalltown Boy AI video generator with no watermark?
Not here. Each video costs credits because it runs on a paid video model. New accounts may get 5 signup credits, enough for a short Seedance 2.0 Mini test but not the full 24 seconds. Free-account videos carry a moving watermark; any paid order unlocks the clean original of the same video.
What AI model makes the Smalltown Boy trend? Is it Higgsfield Genjutsu, Seedance or Kling?
There isn't one official model: TikTok versions were made with motion-transfer tools such as Higgsfield's Genjutsu, Kling Motion Control and Seedance. This generator turns each photo into a character sheet with GPT Image 2.5, then uses Seedance reference-to-video (2.5 for the full 24 s; 2.0 Mini, Fast or 2.0 for shorter clips) or MiniMax H3.
How do I make that Smalltown Boy AI video everyone is posting on TikTok?
Upload two clear photos (the first screamer in Photo 1), choose 24 seconds on Seedance 2.5 for the whole routine, check the credit cost and generate. You get a vertical 9:16 MP4 with the song already on it, ready to post.
Does the Smalltown Boy AI video keep the original song automatically?
Yes. The video is rendered silent and the reference clip's “Smalltown Boy” audio is mixed back in afterwards. The recording belongs to its rights holders, and platforms may mute or restrict it.
How much does it cost to make a Smalltown Boy AI video?
The full 24 seconds on Seedance 2.5 costs 190 credits at 480p, 422 at 720p or 746 at 1080p. Short clips start at 8 credits (Seedance 2.0 Mini, 5 s, 480p). Both character sheets are included, and the exact cost shows before you generate. Failed jobs are refunded automatically.
Can I make a 10-second or 15-second Smalltown Boy AI video for TikTok?
Yes, and 5 seconds too, on any model. Short versions are the opening of the routine: 10 s covers the first scream and the start of the second, 15 s both screams and the first second of the dance. Choose 24 s for the full duo dance and ending.
What is a Smalltown Boy AI video?
An AI remake of a viral dance clip set to Bronski Beat's “Smalltown Boy”: two people take turns screaming into an ultra-wide lens, then dance together and lunge at the camera. This generator puts the two people from your photos into that routine.
Who screams first, and who dances on the left?
Photo 1 screams first (0–7 s) and dances on the viewer's left; Photo 2 screams second (7–14 s) and dances on the right. Roles follow the upload order, not gender, so swap the photos to swap them.
Can I make the full 24-second Smalltown Boy AI video?
Yes, on Seedance 2.5: both screams, the courtyard hand-jive, the low-angle shot, the freeze and the final lunge. The other models stop at 15 seconds.
Can I make the Smalltown Boy AI video with my cats or dogs?
No. This template is built for photos of people, and pet versions are not supported.
Do I need a prompt or a reference video?
No. The reference clip, its cuts and the role instructions are built in. You only choose two photos, a length, a model, a resolution and, if you like, a background.
Will the AI keep our faces, hair and outfits through the dance?
That is what the character sheets are for: faces come from your photos, hair, back view, shoes and outfits from the sheets. Results still vary; a clear front-facing photo that shows the clothes helps most, and Seedance 2.5 holds faces best.
Can I change the background of the Smalltown Boy video?
Yes. Keep the original courtyard, pick the red fence garden or the living room, or upload a picture of your own; the camera moves stay the same and the scene replaces the courtyard in every shot.
What happens if my photos are rejected or the video fails?
With Seedance 2.5 the provider reviews the photos first, and some are refused by its content checks. Any failed or rejected job refunds its credits automatically. Try a different, clearer photo or another model.
Who's screaming first in your duo?
Pick two photos and get a vertical Smalltown Boy AI video with the song already on it.
More two-photo templates: Raindance AI · Hotel Lobby AI · Alors on Danse AI · Rumpelstiltskin AI