How background removal works
Upload a JPG, PNG or WebP and the tool sends it to the U2-Net segmentation model, which predicts for every pixel whether it belongs to the main subject or to the background. The prediction becomes an alpha mask, the mask is cleaned up, and the result is written as a PNG in which the background pixels are fully transparent.
There is nothing to configure: the subject is detected automatically, and the original pixels of the subject are kept untouched. A transparent PNG can then be placed on any colour, gradient or scene. To flatten it onto white for a marketplace listing, run the result through PNG to JPG.
What to expect from the result
U2-Net was trained on salient-object photos, so it works best when one clear subject stands out from its surroundings: a person, an animal, a product on a table, a car. Fine structures such as hair, fur and glass edges are approximated rather than cut pixel by pixel, and low-contrast scenes where the subject blends into the background can leave halos or holes.
Sharp, well-lit photos of a single subject give the cleanest cut-outs. Results are returned at the original resolution up to 8000 pixels per side and 40 megapixels; larger images are rejected with a message rather than silently downscaled.
Typical uses
Online sellers turn phone photos into catalogue images with a clean white background. Recruiters, gamers and forum users make avatars and profile pictures without a distracting room behind them. Designers pull a product or a person out of a photo to compose banners, thumbnails and presentation slides.
Teachers and marketers cut out objects for worksheets and collages, and anyone can remove a busy background before printing a photo on a mug or a card.
After removing the background
The transparent PNG is usually larger than the source JPG because PNG is lossless. Compress Image reduces it for the web without touching the transparency. If the platform you upload to does not accept transparency, convert to JPG, which fills the transparent area with white.
Photos from phones often carry location data in their metadata; the cut-out PNG does not include the original EXIF block. For files you keep in the original format, Remove EXIF strips it.