Image
Remove Background
Two AI modes: fast portrait cutout or general any-subject removal — runs in your browser, never uploads your image.
Loading tool…
Most background removers send your photo to a cloud server — your image travels the network and touches someone else's infrastructure. This tool does none of that. The AI models download once, cache permanently in your browser, and every removal runs entirely on your own hardware. Two modes are available: People & portraits uses MODNet (Apache-2.0, ~25 MB), a compact portrait matting model that is fast on WebGPU and produces clean edges on faces and hair. Any subject uses Silueta (Apache-2.0, ~43 MB), a compressed U-2-Net variant optimised for salient object detection on products, graphics, and objects — it runs on WASM and loads only when you select that mode, so portrait users never fetch it. The output is always a transparent PNG.
Background Removal by Subject
Use-case-specific tips and the right AI mode pre-selected for your subject.
Related Tools
Related Guides
How to Use
- 1
Choose a mode: "People & portraits" (fast, ~25 MB) for photos of people, or "Any subject" (~43 MB) for products, objects, and graphics.
- 2
Drop a PNG, JPEG, or WebP image onto the upload area, or click to browse.
- 3
On first use with a given mode, the AI model is downloaded and cached. People mode downloads ~25 MB; Any subject mode downloads ~43 MB. The progress bar tracks the download. Models are cached permanently — no re-download on future visits.
- 4
Preview the result on a checkerboard background to confirm the cutout, toggle between Before and After, then download the transparent PNG.
Frequently Asked Questions
Is my image ever uploaded to a server?
No. Everything runs in your browser. The AI models are downloaded once and cached; your image is processed locally and never leaves your device. Open your browser's Network tab while using it — you will see no upload request.
What is the difference between the two modes?
"People & portraits" uses MODNet (Apache-2.0, ~25 MB), optimised for matting people, faces, and hair. It loads fast and uses WebGPU on supported hardware. "Any subject" uses Silueta (Apache-2.0, ~43 MB), a compressed U-2-Net model better suited to products, objects, logos, and graphics. It runs on WASM — quality is good on subjects with clear separation from the background, though less precise on fine details like hair compared to portrait mode.
Does selecting "Any subject" download the model immediately?
No — the Silueta model (~43 MB) is only downloaded the first time you actually click "Remove Background" in Any subject mode. Switching the tab alone does not trigger a download. Portrait users who never use Any subject mode will never fetch it.
Why does the first removal take longer?
The first run with a given mode downloads the AI model from a public model repository. A progress bar tracks the download. After that the model is cached in your browser indefinitely — every subsequent removal starts immediately with no re-download.
What is the output format?
The output is always a transparent PNG. PNG is the only widely supported format that preserves the alpha channel (transparency) needed for a clean subject cutout. You can convert the PNG to WebP using the Image Converter tool if you need a smaller web-ready version.