Background Remover

No — the background remover works on your own device. The AI segmentation model and its runtime are downloaded to your browser once, and every image is processed on your own device with WebAssembly. Your photo is never sent anywhere, logged or stored — closing the tab clears everything. That is different from most online background removers, which upload your image to their servers to process it.

A PNG with the background removed. By default the background is transparent, so you can drop the cut-out onto any design. You can also choose a solid white, black or custom-colour background before you download. The output keeps the original resolution of your photo.

The background remover uses a compact open-source neural network (U²-Net) that finds the main subject and separates it from the background. It works best on a clear foreground subject — a person, product, animal or object — against a reasonably distinct background. Fine detail like loose hair or thin, wispy edges is the hardest case and may not be perfect. Because it runs a small model on your device rather than a large one on a server, it favours speed and privacy over the last few percent of edge accuracy.

The first time you use the tool it downloads about 15 MB of model and runtime files, then runs the model in your browser. After that the files are cached, so later images are quicker. Larger photos and slower devices take a little longer because all the computation happens locally.

Any image your browser can open — JPG, PNG and WebP are all fine. For iPhone HEIC photos, convert them first with the HEIC to JPG converter, then remove the background. The download is always a PNG, because PNG is the format that supports transparency.