Remove Background

Cut a person out of a photo, entirely in your browser. The image never leaves your device. Built on a lightweight open-source segmentation model, so it works best when a person is clearly the main subject.

Need it with no sign-up, no watermark, and no limits? Try Remove Image Background Free.

Heads up: this uses MODNet, a portrait-matting model tuned for people, not general objects or products. Portraits and selfies work best. The model and the WebAssembly runtime it needs are about 18 MB combined and download once per browser session, then stay cached.

Click to choose a photo, drag and drop it here, or paste with Ctrl+V

Works best with one person, clearly in frame. Select multiple photos to batch process.

Your cutout will appear here once processed.

How it works

The heavy lifting is done by MODNet, an open-source portrait-matting network, run through ONNX Runtime Web (Microsoft's WebAssembly build of the ONNX inference engine) and served entirely from this site's own files. Unlike a plain segmentation model, MODNet is trained to predict a real alpha matte, a per-pixel opacity value, rather than a blunt "person or background" guess, so hair and soft edges come out closer to true transparency instead of a hard cutout line. The image is first resized so its longer side lands near 512px, since that's the resolution the model was trained at; feeding it a bigger photo doesn't sharpen the result, just slows processing down.

Even a real alpha matte has some noise at its edges, so the tool still does a light cleanup pass: a small box blur smooths resizing artifacts, then the alpha curve gets a gentle smoothstep remap so background haze near zero snaps fully transparent and confident subject pixels snap fully opaque, while the soft transition in between (hair, blur, motion) is preserved rather than flattened.

The network story is verifiable in DevTools: on your first run, the Network tab shows the model and runtime files loading from this site, about 18 MB combined. After that, process as many photos as you like and watch: no request ever carries your image. The photo is read into memory, matted, and composited entirely in this tab.

Common questions

Why does it only work well on people?

It runs MODNet, an open-source portrait-matting model that was trained specifically to separate people from backgrounds. Products, pets, and objects aren't what it learned, so those cutouts will be rough. Portraits with the person clearly in frame work best.

Is my photo uploaded anywhere?

No. The model runs inside your browser using ONNX Runtime Web, so your photo never leaves your device. The only thing downloaded is the model and runtime itself, about 18 MB combined, once per browser session, and it's cached after that.

Why do the edges look soft or rough?

The model works at a capped internal resolution (around 512px on the longer side), so very fine detail like flyaway hairs can still come out soft on large photos. You'll get the cleanest edges from sharp photos with good contrast between the person and the background.

Can I replace the background instead of making it transparent?

Yes, pick a solid color under "Result background" before processing. The download is always a PNG, so if you keep it transparent you can layer the cutout over anything in another editor.