About BGRemover

BGRemover is part of a small collection of free online tools built by an independent developer. The goal of the project is simple: take tasks that usually require paid software, accounts, or risky file uploads, and make them available to everyone — free of charge and without compromising privacy.

Why we built this

Removing a background from a photo should not require a subscription, a watermark on your own image, or uploading personal photos to somebody else's server. Most online background removers charge for HD downloads, cap free usage, or quietly turn your images into training data.

We took a different route. Modern browsers are powerful enough to run real neural networks locally through WebAssembly. So BGRemover downloads a compact open-source AI model to your device once, and from that moment on, every image you process stays entirely on your device. No upload queue, no server farm, no privacy trade-off.

Free foreverNo credits, no watermarks, no download limits

Private by designImages are processed on-device and never uploaded

Open source insideBuilt on U²-Net and ONNX Runtime Web

How it works under the hood

U²-Net — an open-source deep neural network for salient object detection (Apache License 2.0), compact enough to run in a browser tab.
ONNX Runtime Web — Microsoft's MIT-licensed inference engine that executes the model in WebAssembly on your own CPU.
Pure front-end stack — everything else is plain HTML, CSS, and JavaScript. There is no image-processing backend at all.

Because the heavy lifting happens on your hardware, we can offer unlimited HD exports, batch processing, and background replacement without charging a cent.

What BGRemover offers

Automatic AI background removal for people, products, pets, and objects
Transparent PNG export in full resolution (up to 4096 px)
Background replacement: solid colors, blur, or your own photo
Batch processing of up to 30 images with one-click ZIP download
Interface in 9 languages, free for personal and commercial use

Open source credits

This site stands on the shoulders of open-source projects, used in compliance with their licenses: U²-Net (Apache-2.0), ONNX Runtime Web (MIT), and JSZip (MIT). We are grateful to their authors and communities.

Get in touch

Feedback, bug reports, and feature requests are always welcome — they directly shape what we build next. Reach us any time at guichenli86@gmail.com, or use the form-free details on our contact page.