Click any color in an image to make it transparent with adjustable tolerance and anti-aliasing.
A background or matte color is removed by clicking that color directly in the image. The tolerance control determines how far nearby colors may differ from the sampled color and still become transparent, while anti-aliasing can soften the resulting boundary. The important surprise is that this is color-based removal, not subject detection. A matching color inside the object can disappear along with the background, even when the matching areas are separate.
A concrete input and expected output from the current implementation.
Input
Upload a 2 by 2 image with rows [#FFFFFF, #FF0000] and [#000000, #FFFFFF]. Click #FFFFFF, set tolerance to 0, and turn anti-aliasing off.
Expected output
The resulting rows are [transparent, #FF0000] and [#000000, transparent]. Two pixels are fully transparent, and the red and black pixels remain fully opaque.
At zero tolerance, only pixels exactly matching the selected white are removed. The image contains two white pixels, so exactly two of its four pixels become transparent.
Clicking the image selects the color at that pixel as the removal target. With a low tolerance, the tool affects colors that closely match the selection. Raising tolerance admits a wider range of neighboring colors, which can help with compressed, shaded, or uneven backgrounds but can also remove similar colors belonging to the foreground.
Matched pixels lose opacity instead of being painted another visible color. Anti-aliasing softens the transition around the removed region so curved or diagonal boundaries appear less jagged. This can leave partially transparent edge pixels, which may look different when the result is placed over dark, light, or strongly colored backgrounds.
There is no single correct setting because tolerance depends on variation in the background and similarity between background and subject colors. Start at the lowest value that removes the clicked color, then increase it gradually while watching fine foreground details. If parts of the subject begin disappearing, the selected color and image may not support clean removal by tolerance alone.
No. Making pixels transparent removes their visible contribution; it does not reconstruct content that was covered when the source image was created. The result exposes whatever background is later placed beneath the image. If the selected color overlaps important subject pixels, those pixels cannot be recovered as hidden detail by this operation.
The example and behavioral notes were checked against the browser implementation. Standards and primary references below define the relevant format, formula, or platform behavior.
Pick colors visually and convert between HEX, RGB, HSL, and CMYK formats.
Preview how screen colors translate to print with CMYK simulation and proofing.
Preview any Google Font at various sizes, weights, and with custom sample text.
Create linear, radial, and conic CSS gradients with a visual editor and code output.
Generate harmonious color palettes from a base color using color theory rules.
Simulate how colors appear to people with various types of color vision deficiency.