How to use
- Load your image, then open the Background section and click Enable Click-to-Remove.
- Set the Tolerance slider, then click the color you want to remove. The image updates right away.
- Undo if too much disappeared, change the tolerance, and click again. Download when you are happy.
This is a color-based remover, not an object detector. When you click a pixel, the tool reads its color and then makes every pixel in the image transparent whose color is close to it. Closeness is measured as the distance between the two colors in red, green and blue values, and the Tolerance slider (1 to 100) sets how far a color may be from the one you clicked and still be removed.
That makes it a good fit for images with a flat or evenly colored background, such as a product on white, a logo on a solid color or a sprite on a green screen. It is not a good fit for photos where the background has gradients, shadows or the same color as part of the subject, because the tool has no idea what the subject is. Pixels elsewhere in the picture that share the clicked color are removed too, not only the ones touching the edge.
For uneven backgrounds you can click several times. Each click is a separate step in the history, so you can use undo and redo in the top bar to remove one pass at a time. Increase the tolerance a little for soft or noisy backgrounds, and lower it if parts of the subject start to disappear.
The edited image is kept with transparent pixels, which PNG and WebP can store. Everything runs in your browser on a canvas, and the image is not uploaded.
FAQ
Does it detect people or objects automatically?
No. It removes colors you click, not objects. It uses no machine learning model.
What kind of images work best?
Images with a single flat background color, such as scans, logos, icons and sprites.
Why did part of my subject disappear?
Those pixels were within the tolerance of the color you clicked. Undo, lower the tolerance and click again.
Is transparency kept?
The edited image holds the transparent pixels. Formats that support alpha, such as PNG and WebP, keep them. JPG cannot store transparency and fills those areas with white.
Is my image uploaded?
No. The color matching runs in your browser, and nothing is sent to a server.