Drive multiple characters in one scene

Prepare a separate reference image for each role, then use one shared driving video to guide a duet or group interaction. Clear subject visibility and manageable overlaps help the model track the scene.

Upload several role images and keep each one labeled in the Studio.
Use your own driving video or pick a prepared motion reference.
Preview the selected inputs and generate the composed video.

Open Studio with this workflow

Only upload media you have permission to use. Model and provider support can vary by workflow.

Example outputMulti-character · shared motion

Add character references

Upload one image per role. The Studio assigns an ID to each image so you can review the set before generation.

Choose the driving clip

Select a prepared reference clip or upload a short video with clearly visible movement.

Create and review

The selected mode and input media are sent with the task. Review the result in your generation history.