Images in your designs
Generating images, editing them in place, uploading your own, and replacing an asset across every screen.
Last updated
Generated images
When a design needs imagery, the AI generates it. You can steer it:
Regenerate the hero image: a sunlit open-plan office, wide shot, warm tones, no people.
Describe subject, setting, composition, lighting and mood. A bare noun ("a laptop") produces a bare result.
Image generation is charged per image, at a fraction of the cost of a screen. See How credits work.
Editing an existing image
You can change an image in place, keeping everything else pixel-identical:
In the hero image, change the text on the sign to "Ledgerly". Recolour this illustration to match the new theme.
This is the right tool for changing text inside an image, recolouring, restyling, or adding and removing a single element. Regenerating from scratch would change everything else too.
Uploading your own
Upload through the prompt box or the project's asset library, then reference it:
Use the photo I just uploaded as the hero image on the landing page.
Replacing an asset everywhere
To swap a logo or product shot across the entire project, ask for exactly that:
Replace the old logo with the new one on every screen.
This runs as a single deterministic operation across all screens rather than a per-screen edit â it's the only reliable way to be sure you didn't miss one.
Aspect ratio
State it if it matters â "16:9 hero", "square avatar". Otherwise you may get a shape that doesn't fit the slot.
If images are broken or stuck loading
Was this article helpful?
Related articles
Editing with chat
How to phrase AI edit requests so you get the change you meant, on the screens you meant.
Editing a screen
The three ways to change a design â chat, edit mode and theming â and when each is the right tool.
Images are broken, missing or stuck loading
Placeholder boxes, permanently loading images and images that vanish after export.