Cutout for use in any program
- Input
- fundo transparente
- Expected output
- PNG
Lossless, with alpha, accepted everywhere.
Carregando ferramenta… / Loading tool…
make transparent png from photo
Leave the background on Transparent, which is the default, and download. The checkerboard shown in the preview is just the visual convention for "empty": it is not part of the file.
Lossless, with alpha, accepted everywhere.
Keeps alpha; check that your editor opens WebP.
JPG has no alpha channel.
A mix of 50% red and 50% white.
That is why the edge looks different depending on the background.
No. The photo is read and processed in your browser, inside a Web Worker. The internet is only used to download the cutout model (from Hugging Face) and the ONNX Runtime engine (from jsDelivr), once each. You can verify it in your browser's Network tab: after the model files are downloaded, no request carries your image.
The checkerboard is the visual convention for transparent areas and exists only on screen. The downloaded file has alpha set to zero in those areas, with no checkerboard drawn.
Open Edge refinement, keep the option to remove the old background color from the edge on, and use Shrink with a small value such as -2. That eats the fringe of pixels that still carries the original background color.
Yes. The model runs once per photo; changing the color, gradient, blur or background image only recomposes the image, and it is instant.
Your photo never leaves your device. The cutout runs here, in your browser. Only the model (and the engine that runs it) is downloaded, once; no image is sent to any server.
MODNet, built for people and portraits. Small and quick; weaker on products, animals and objects.
All processing happens in your browser. The only things fetched from the internet are the cutout model (Hugging Face) and the ONNX Runtime inference engine (jsDelivr), downloaded once and checked by size and SHA-256. Your photos are never uploaded.