Image background remover
Choose a picture, click the background color in the preview, adjust the tolerance and download a PNG with a transparent background. The method is color-based, not AI, so it works best on plain, evenly colored backgrounds. The file stays on your device.
Runs in your browser: nothing you enter is sent anywhere.
How to use it
- Choose a picture. The color of its top-left corner is picked as the background automatically.
- Click the background in the preview to pick a different color, or use the color box.
- Raise Tolerance until the background is gone but the object is intact. Choose whether to remove only the background touching the edges or every pixel of that color.
- Add a little Edge softness (1 or 2 px) for a smoother cut, then download the PNG.
How it works
Each pixel is compared with the picked color as a point in red-green-blue space. The longest possible distance is the one from black to white, about 441.7; a tolerance of 20 % means pixels within about 88 units count as background. In edge mode a flood fill starts at every border pixel that matches and spreads to its four neighbors. Edge softness blurs the transparency map with a box of that radius so the cut fades instead of stepping.
Worked example
A product photo on a white sheet: pick the white, set the tolerance around 15 to 25 % and keep the edge mode. The sheet disappears, while white parts of the product that are surrounded by other colors stay. If a gray shadow remains, raise the tolerance a little or accept it as part of the object.
Tips
- Photos with soft shadows or gradients need a higher tolerance, which can start to eat into similar colors of the object.
- Pictures with a flat background, such as logos, icons and screenshots, give the cleanest results.
- JPG compression adds noise around edges; a PNG source gives sharper cuts.
Frequently asked questions
How does this remove the background without AI?
It compares every pixel with the color you picked. A pixel counts as background when its distance from that color in red, green and blue is within the tolerance. The tolerance is a percentage of the longest possible distance (black to white). Pixels that match become transparent.
What is the difference between the two removal modes?
Around the edges starts at the picture border and spreads only through matching pixels that touch each other, so a white shirt inside a white background stays. This color everywhere removes every matching pixel, including enclosed areas such as the holes in a letter or a ring.
Why are there jagged or colored edges?
Edge pixels are a mix of the object and the background, so they may be only partly matching. Raise the tolerance a little, or use Edge softness to blend the cut. For hair, fur or glass on a complicated background, a color-based method is not enough.
What does the saved file look like?
A PNG at the original resolution with an alpha channel. The preview is calculated on a smaller copy (up to 800 pixels on the long side) for speed, and the download is calculated again on the full-size picture.