The short version
Start with a sharp, well-lit photo where the subject stands out clearly from what is behind it. Run it through the remover, turn on auto-crop if you want a tight edge, adjust the edge strictness if the outline looks soft or harsh, and save as PNG so that the transparency is kept.
How it works
The remover uses an AI model that has learned what the main subject of a photo usually looks like. For every pixel it estimates how likely it is to belong to the subject and how likely to belong to the background, and it uses that to build a mask. The mask is applied to your photo to make the background transparent.
The model is downloaded to your browser the first time you use the tool, which is why the first photo takes longer. After that, the photo is processed on your own device, so it is not uploaded to a server.
What gives a clean result
- Contrast between subject and background. A person in front of a plain wall is easy. A person in a green jumper in front of a green hedge is hard.
- Good, even light. Dark or very shadowy photos give the model less to work with.
- A sharp subject. Blur at the edges becomes uncertainty in the mask.
- A subject that fills the frame. Tiny subjects in a large scene have less detail for the model to judge.
- A simple background. If you are about to take a photo for a cut-out, shoot against a plain wall or sheet.
The hard cases
Some things are difficult for every automatic tool. Fine strands of hair, fur and lace have no clean edge, because the background shows through the gaps. Glass, water and other transparent objects are neither subject nor background. Shadows on the floor can be treated as part of the subject or removed with the background, depending on the photo. If the cut-out of such a subject is not perfect, a slightly lower edge strictness often keeps more of the soft detail.
Edge strictness and auto-crop
Edge strictness controls how decisive the tool is about borderline pixels. At a low setting it keeps semi-transparent edges, which looks natural around hair and soft clothing. At a high setting, every pixel that is not confidently part of the subject is removed completely, which gives crisp edges that suit products and logos. If you see a faint outline of the old background, increase the strictness a little.
Auto-crop trims away the empty transparent border so that the picture is cut tight to the subject, which is handy for stickers, profile pictures and product listings.
Avoiding halos on a new background
A common problem is a thin light or dark fringe around the subject when you place it on a new background. It happens because the edge pixels still carry a little colour from the original background. You can reduce it by raising the edge strictness, by putting the cut-out on a background of a similar tone to the original, or by choosing a photo where the old background was close in colour to the new one.
What to do with the result
Save the result as a PNG, which keeps transparency. A JPG cannot be transparent and will fill the empty area with white or black. You can drop the PNG onto a new background in any editor, use it as a product photo on plain white, make a profile picture, or turn it into a printable sticker with the Sticker Maker.
Try it yourself
Last updated 2 November 2026. These guides are general information written by the Cwtch.es team. Settings and platform limits change, so check the current details for any service you rely on.