Background Remover
Remove the background from a photo automatically with an in-browser AI model — export a transparent PNG or drop in a new background. Your image never leaves your device.
Drag & drop an image here, or
JPG, PNG, WebP, GIF, BMP · or paste with Ctrl/⌘+V
Your image is processed locally and never uploaded. The AI model (a few MB) is downloaded once from a CDN the first time you use this tool.
Works best on photos with a clear subject — people, products, animals. Very fine detail (loose hair strands) or busy backgrounds can be tricky.
Loading the model…
This can take a few seconds — the model runs entirely in your browser.
What is the background remover?
A background remover automatically separates the main subject of a photo — a person, product or animal — from everything behind it, so you can place the subject on a transparent background, a solid colour, or an entirely new scene. What makes this tool unusual is that it does the whole job inside your browser using a machine-learning segmentation model compiled to WebAssembly. There's no server doing the work and no API call: the image itself is never uploaded. The one thing that is fetched is the AI model file (a few megabytes), which downloads once from a CDN the first time you use the tool and is then cached by your browser. After that, removing a background is a local, private operation. It's ideal for product shots, profile pictures, simple graphics and any time you need a clean cut-out without paying for a subscription service or handing your photos to a website.
How it works
Add a single image by dragging it in, choosing a file, or pasting from the clipboard. The first time, the tool downloads and initialises the model, showing a clear progress indicator — this takes a few seconds and is much longer than the other image tools, which is normal for on-device AI. The model then analyses the picture and produces a cut-out of the subject with a transparent background. You'll see a before/after comparison with a slider so you can check the edges. From there you choose the output: keep it transparent as a PNG, place the subject on a solid colour with the colour picker, or upload your own background image to composite behind it. When you're happy, download the result — a transparent PNG by default, or JPG/WebP when you've added an opaque background. The heavy computation runs off the main thread where the library supports it, so the page stays responsive, and if the model can't be fetched (for example, a network problem) or an image is unusually large and times out, the tool shows a clear error rather than hanging.
Worked example
You photographed a product on a cluttered desk and need it on a clean white background for a shop listing. You load the photo; the model downloads on first use and then processes the image in a few seconds. The before/after slider shows the desk gone and the product cut out. You switch the background to a solid white, check that the edges look clean, and download a JPG ready for the listing. Next you want the same product as a transparent PNG to drop onto a coloured banner, so you switch the background back to transparent and download again — no re-processing needed.
Frequently asked questions
Is my image uploaded to a server?
No. The image is processed locally in your browser. Only the AI model file is downloaded once from a CDN; your photo itself never leaves your device.
Why does it take a few seconds?
The first run downloads a multi-megabyte AI model and then runs it on your device. That's slower than simple canvas tools, but it means your image stays private. Later images are quicker once the model is cached.
What kinds of images work best?
Photos with a clear subject — people, products, animals — against a reasonably distinct background. Very fine detail like loose hair strands, or busy/low-contrast backgrounds, can produce rougher edges.
Can I add a new background?
Yes — keep it transparent (PNG), fill it with a solid colour, or upload your own background image to composite the subject onto.
What format is the output?
A transparent PNG by default. If you add a solid colour or image background, you can also export JPG or WebP since transparency is no longer needed.
Does it work offline?
After the model has been downloaded and cached once, processing is local. The very first use needs an internet connection to fetch the model.
What if the model fails to load?
If the model can't be fetched (e.g. a network issue) or a very large image times out, the tool shows a clear error message so you can retry with a connection or a smaller image.
Related tools
- Image Format ConverterBatch convert images between JPG, PNG, WebP and AVIF — with quality control, resizing and ZIP download, entirely in your browser.
- JPG to PNG ConverterConvert JPG (JPEG) images to lossless PNG — privately in your browser, one file or a whole batch.
- PNG to JPG ConverterConvert PNG images to smaller JPG (JPEG) files, choosing a background colour to replace transparency.
- WebP to JPG ConverterConvert WebP images to universally-compatible JPG (JPEG), with a background colour for any transparency.
- JPG to WebP ConverterConvert JPG (JPEG) images to WebP to make them smaller and faster for the web — in your browser.
- Image ResizerResize images by exact pixels, a percentage, or a social-media preset — with aspect-ratio lock and batch ZIP output, all in your browser.
- Image CropperCrop an image with a draggable box and aspect-ratio presets, rotate and zoom, and export at full resolution — privately in your browser.