Two photos, two fixed roles
Photo order decides who performs each role. The wide dance shots place Photo 1 on your left and Photo 2 on your right. Their background positions during the close-ups are different, as shown below. All left and right positions are from the viewer’s perspective.
| Reference time | Photo 1 | Photo 2 |
|---|---|---|
| About 0–7s | First screaming close-up | Dances in the background on the left |
| About 7–14s | Dances in the background on the right | Second screaming close-up |
| About 14–24s | Left performer in the shared shots | Right performer in the shared shots |
These are the roles in the motion reference. AI output can vary in timing, expressions, hands and identity details. Use a sharp photo with one visible face per image; avoid group photos, covered faces and heavy filters.
Choose the part of the routine you want
Every shorter option uses the opening of the same reference. Choose 24 seconds for both close-ups and the complete dance sequence.
| Length | What the reference contains | Available models |
|---|---|---|
| 5 seconds | The opening close-up, with the second person dancing behind | Seedance 2.5, Seedance 2.0, Seedance 2.0 Fast, Seedance 2.0 Mini |
| 10 seconds | The first close-up and the beginning of the second | Seedance 2.5, Seedance 2.0, Seedance 2.0 Fast, Seedance 2.0 Mini |
| 15 seconds | Both close-ups and the start of the shared dance | Seedance 2.5, Seedance 2.0, Seedance 2.0 Fast, Seedance 2.0 Mini |
| 24 seconds | The full routine, including the shared dance and final reach | Seedance 2.5 |
All videos are vertical 9:16 MP4s, suitable for a TikTok, Reels or Shorts post. Supported resolutions: Seedance 2.5: 480p, 720p, 1080p; Seedance 2.0: 480p, 720p, 1080p; Seedance 2.0 Fast: 480p, 720p; Seedance 2.0 Mini: 480p, 720p.
Change the setting, keep the routine
Keep the reference courtyard, move the performance to a red-fence scene or living room, or upload your own background image. Your two people follow the same reference routine in the selected setting. Use a clear location photo without foreground people for an easier-to-read scene.
The background is generated from your image and may change with the camera angle. It is not an exact photographic composite. For a different clip and more editing instructions, try the background video editor.
Hear the song in your downloaded video
The finished MP4 includes the selected reference clip’s audio, featuring “Smalltown Boy” by Bronski Beat. You do not need to add the song separately to hear it in your download. Shorter videos use the matching opening audio.
Audio being included does not grant a music license. Check the rights needed for your intended use and the platform where you publish. The service is not affiliated with Bronski Beat or EachLabs.
Credits, free access and downloads
Opening clips start at 18 credits. The complete 24-second routine on Seedance 2.5 at 480p costs 200 credits. Every price below covers preparation of both characters and the finished video, including its audio. Background choices do not add a separate fee. A dash means that combination is unavailable.
| Model | 5s | 10s | 15s | 24s |
|---|---|---|---|---|
| Seedance 2.5 · 480p | 53 | 91 | 130 | 200 |
| Seedance 2.5 · 720p | 101 | 188 | 275 | 432 |
| Seedance 2.5 · 1080p | 169 | 323 | 478 | 756 |
| Seedance 2.0 · 480p | 41 | 68 | 95 | — |
| Seedance 2.0 · 720p | 72 | 129 | 187 | — |
| Seedance 2.0 · 1080p | 159 | 304 | 449 | — |
| Seedance 2.0 Fast · 480p | 30 | 46 | 62 | — |
| Seedance 2.0 Fast · 720p | 48 | 83 | 117 | — |
| Seedance 2.0 Mini · 480p | 18 | 23 | 27 | — |
| Seedance 2.0 Mini · 720p | 23 | 33 | 42 | — |
The generator shows the exact price for your selection before you start. See the live signup-credit and watermark rules in the FAQ below, and compare credit plans. If preparation, input review, video generation or final processing fails, the charged credits are returned. A completed video with imperfect faces or movement is not an automatic failed-job refund.
Where the motion reference comes from
The preview and role frames come from the Smalltown Boy workflow published by EachLabs. They show the intended moves and camera changes; they are not generated results from this tool. Your uploaded photos and prepared character references remain private inputs.
For another two-person routine, try Alors on Danse. To replace people in your own clip, use the duo video editor.
Last updated: 2026-10-10