Motion Workflows

A motion workflow is the process of creating a video by combining an image and a reference video. On the canvas, you can build these workflows visually by connecting nodes.

Basic Video Workflow

1. 1Generate or select an image of your AI identity
2. 2Upload a reference video to an upload node
3. 3Connect the image source to the video generation node's image input
4. 4Connect the reference video to the video generation node's video input
5. 5Optionally add a prompt
6. 6Generate

Image-to-Video Pipeline

A common workflow is to generate an image first, then turn it into a video. Connect an image generation node's output to a video generation node's image input. This lets you create a photo and then animate it in one continuous pipeline.

Using Different Reference Videos

You can connect different reference videos to the same video generation node to see how your identity looks performing different motions. Or create multiple video generation nodes, each with a different reference video, to generate several variations.