Expand Image Canvas
Grow the frame around a photo and choose where it sits.
Paint over something small and have it filled in.
Drop an image here, or click to choose one
Paint over something small · the smaller it is, the better this works
The Object Eraser fills a painted mask by diffusion — repeatedly averaging inwards from the edges, so colour and gradient flow in from all sides. On a blemish, a wire against sky or a small object on a plain surface it is invisible. On texture it smudges, and the page says so rather than letting you find out.
Removing something from a photograph means inventing what was behind it, and how well that can be done depends entirely on how complicated the surroundings are.
The method here is diffusion inpainting: the masked area is filled by repeatedly averaging from its edges inwards, so colour and gradient flow in from all sides. On a blemish, a stray wire against a sky, a sensor spot, a small object on a plain surface, the result is indistinguishable. On something large, or against detailed texture that has to be recreated, it produces a smudge — and the page says so rather than letting you discover it. Brush size, iteration count and edge feathering are all adjustable, and the mask can be undone.
A classical algorithm that runs instantly with no model download, and reconstructs smooth areas convincingly.
Size and softness, with the mask shown as you paint so you can see exactly what will be replaced.
More passes give a smoother fill; the count is yours to set rather than fixed.
Paint, run, and undo freely — the original is never modified until you download.
The page states where diffusion works and where it smudges, which saves a lot of wasted attempts.
The smaller the thing you want gone, the better this works.
Cover the object and a little of its edge — a mask that stops short leaves a rim behind.
The area is filled from its edges inwards over several passes.
Add to the mask and run again, or undo and start over.
Small ones. A blemish, a wire, a sensor spot, a sign on a plain wall. Anything more than a few percent of the frame, or against complex texture, will smudge rather than fill.
Diffusion inpainting: the masked pixels are repeatedly averaged from their neighbours, so colour and gradient flow in from the edges. It reconstructs smooth areas well and texture not at all.
Because the surroundings had texture the fill cannot recreate. Diffusion produces smooth colour, so it is invisible against a sky and obvious against gravel.
No. Generative inpainting would need a large model download and a lot of compute. This is a classical algorithm that runs instantly and is honest about its range.
What people who use the Object Eraser think of it. Reviews come from visitors and are posted as written.
No ratings yet
Yours would be the first.
0 reviews
Newest first
No reviews yet
Be the first to rate the Object Eraser. It takes half a minute and no account.
Write the first reviewGrow the frame around a photo and choose where it sits.
Put a screenshot on a gradient with shadow and rounded corners.
Fix a tilted horizon with a fine angle slider and grid.
Read a QR code from an image or your camera.
Cut a photo into an even grid of tiles and download a ZIP.
Turn a Base64 string or data URI back into an image file.
Turn a clip from a video into an animated GIF.
Resize photos to every Instagram format without cropping.