AI Dance Generator: Make a Photo Dance

Upload a dance video as the motion reference and one photo of your character, and get a new clip of that character doing the dance.

One reference video (3–30 seconds) and one character image per clip.

How this AI dance generator works

The AI dance generator in AdsTurbo is built on Motion Control, which transfers movement from one video onto a different character. You supply two things: a reference video of someone dancing, and a single image of the character you want to see dance. The video decides the moves and the image decides the look. The output is a new clip in which your character performs the choreography from the reference.

Because the moves come from real footage, you are not limited to a fixed list of preset dances. Any routine you can record or have permission to use can become the reference: a trend you filmed yourself, a short choreography from your team, or a product-holding gesture sequence for an ad.

Choosing a dance reference video

  • Format and size: MP4, AVI or MOV, up to 200MB.
  • Length: 3–30 seconds when the clip follows the video framing, or 3–10 seconds when it follows the image framing.
  • One dancer, clearly visible from start to finish. Group routines and clips where the dancer leaves the frame are harder to follow.
  • Continuous movement with little occlusion. Props, furniture or other people passing in front of the dancer reduce accuracy.
  • Very fast footwork and close-up hand moves are the parts most likely to drift, so check them in the result.

Choosing the photo that will dance

To make a photo dance, pick an image where the character is easy to read: JPG, PNG or WebP, up to 10MB, in a common ratio such as 9:16, 16:9, 3:4, 4:3 or 1:1. Framing matters more than resolution. If the reference shows a full-body routine, a full-body character image gives the model legs and feet to animate; a waist-up portrait paired with a full-body dance forces it to invent the rest.

The same character image can be reused in separate generations with different dance videos, which is a simple way to build a series around one model or brand character.

How to make a photo dance in AdsTurbo

  1. Upload the dance video

    Open Motion Control and add your reference dance video as the motion source.

  2. Upload the character image

    Add the image of the character who should perform the dance.

  3. Set resolution and orientation

    Choose 720p or 1080p. Pick Match video to follow the camera angle and framing of the dance clip, or Match image to keep the framing of your photo.

  4. Describe the background if you want

    The optional prompt only describes the background. It does not change the character, the moves or the expression.

  5. Generate and review

    Generation takes about 3–4 minutes on average. Check the face, hands and fast sections before you download or edit the clip.

Music, length and what it does not do

  • The original audio of the reference video is kept in the generated clip. It cannot be replaced or removed inside Motion Control, so choose a reference whose sound you are allowed to use.
  • It does not create choreography from a song or from a text prompt. The dance always comes from the reference video.
  • The clip is as long as the reference allows: up to 30 seconds with video orientation, up to 10 seconds with image orientation.
  • Output is 720p or 1080p. Other sizes need a separate upscaling step afterwards.

Credits depend on length and resolution, and the estimate is shown in the workspace before you generate. As a current example, a 15-second clip uses 315 credits at 720p or 420 credits at 1080p, and credits are refunded if generation fails. Use reference videos and character images you have permission to use, especially when the dancer or the character is a real person.

Frequently asked questions

How does an AI dance generator make a photo dance?

It transfers motion. In AdsTurbo you upload a dance reference video and one character image; the video guides the movement and the image defines the character, and the result is a new clip of that character doing the dance.

Can it generate a dance from music or a text prompt?

No. The moves always come from a reference dance video. The optional prompt only describes the background, and there is no option to create choreography from a song.

How long can the AI dance video be?

The reference video can be 3–30 seconds with video orientation or 3–10 seconds with image orientation, up to 200MB. The generated clip follows the reference.

Does the dance video keep the music?

Yes. The original audio of the reference video is preserved in the result. It cannot be swapped or removed inside Motion Control.

Can I use the same photo with different dances?

Yes. Reuse one character image across separate generations, each with its own reference video, and check character consistency in every result.

Turn a dance video into your character’s dance

Upload a reference dance and a character image, choose 720p or 1080p, and review the clip in the workspace.