AI inpainting

Inpainting means editing only a masked region of an image while the rest stays untouched. You paint a mask over the area you want to change, describe what should be there, and the model regenerates just that region to match the light, texture, and perspective of the surrounding frame. It is the method behind removing objects, filling gaps, and swapping parts of a photo, all from the same mask-and-generate step.

A full image edit reworks the entire frame, so parts you wanted to keep can drift. Inpainting locks everything outside the mask and only regenerates what is inside it, so a face, a product, or a background you were happy with stays pixel-stable while one region changes. That control is the whole point of masking.

Remove something and rebuild the space behind it, add an element that was not there, replace an object with a different one, fix a flaw, or change a detail like a colour or a texture. Because you control the mask and the prompt, the same tool covers removal, replacement, and repair.

Both. Leave the prompt empty and the model fills the mask from the surroundings, which is how removal works. Add a prompt and it generates what you describe inside the mask instead, so you can paint in a new object, a different sky, or a change of clothing while the rest of the frame holds.

That is what inpainting is built to do. The regenerated region is conditioned on the pixels around the mask, so it continues the lighting and texture of the frame rather than sitting on top as a paste. On a hard edge you can feather the mask, rerun, or switch models until the seam disappears.

Yes. Once a frame is inpainted the way you want, you can carry it straight into image-to-video on Morphic and animate the shot, so the still you just fixed becomes the first frame of a clip. Editing and inpainting run on Compose, and the canvas is where you generate and compare.

No. Inpainting runs on Morphic in the browser. Upload an image, mask an area, generate, and export from the same canvas.