In short: The AI panel generates video from a prompt or from an image you already have.

Video generation sits with the other AI tools in the editor, and it works two ways: from a description, or from an image you already have.

Generating from one of your own images is usually the better route, because it starts from something already on brand rather than from nothing.

1

From a prompt

Describe the shot you want. This is the route when you have nothing to start from.

2

From an image

Take an image already in your layout and put it into motion. The quick action carries the image across, so you are not re-selecting it.

3

Set the shape and length

Aspect ratio and duration are set before you generate rather than fixed afterwards, so pick the ratio of the place it is going.

4

Decide about audio

There is a Generate Audio switch, off by default. Leave it off unless you want it, since generated music is rarely what you would have chosen.

5

Close the panel and carry on

Generation takes a few minutes. You do not have to sit and watch it, the panel can be closed and you are notified when the video is ready.

6

Find it again

Generated video is kept in your AI Video library, so you can reuse it without regenerating and spending again.

Generation costs Sparks, and you will be told before you run out rather than after.

Last updated 1 September 2026