Stable Diffusion Inpainting 🎨

Upload an image, paint a mask over the area to replace, and add a text prompt for what you want to see there.

What should appear in the painted area? Describe it below, then start the generation.

Input Image

Upload an image
or
select an example below to start

Brush 40

Output

Advanced options

guidance_scale steps strength seed

negative_prompt

Sampler Schedule

AI inpainting regenerates only the image area you paint over: the rest of the original image is kept pixel for pixel. It works for object removal (paint over the object, describe the background), for replacing a detail, a sky or a piece of clothing, and for extending a scene. Stable Diffusion fills the masked area from your prompt and blends it with the surrounding pixels, so a short, concrete description of what should appear there gives the cleanest result.

Examples

How to use Stable Diffusion inpainting

Upload a JPG, PNG or WebP image, or pick one of the examples.

Paint the mask over the zone to replace; brush size and eraser are in the toolbar.

Describe what should appear in the masked area, in English, as in a normal prompt.

Click Generate, watch the progress, then download the result or repaint another zone.

Limits and questions

What image sizes are accepted?

Any size up to a few megabytes; images larger than 1024 px on a side are resized proportionally before generation, and the result comes back at that working size.

How long does it take?

Jobs run one at a time on one GPU: usually 10 to 40 seconds, more at busy times. You see your position in the queue and a live preview. One submission per IP every 20 seconds.

What is refused?

Adult content and violence are not allowed anywhere on the site: such prompts are refused before generation. Results are kept for a few minutes only and are not published.