CanDoYa
EN

Free AI Background Remover

Runs entirely in your browser - no upload, no sign-up.

Choose a portrait or clear single-subject photo to begin.

Share this tool

What does a background remover do?

A background remover separates the main subject from its surroundings and turns the removed area transparent. This tool runs a portrait-matting model inside your browser, preserves the original pixel dimensions, and exports a PNG. Your photo stays on your device while the public AI model downloads separately.

How to use

  1. 1Choose a suitable image. Drop in a JPEG, PNG, or WebP up to 25 MB and 24 megapixels. Portraits and photos with one clear foreground subject give the model the strongest signal.
  2. 2Run the local AI model. Click Remove background. Wait for the first model download and on-device processing; WebGPU is used when available, with an automatic CPU fallback.
  3. 3Inspect and download. View the result over transparency, white, and dark backgrounds. Check difficult edges, then download the full-dimension transparent PNG.

Who it's for

The tool uses the Apache-2.0 MODNet portrait-matting model in a browser worker. It prefers WebGPU with 16-bit model weights and falls back to WebAssembly on the CPU when WebGPU is missing or fails. The first run downloads roughly 13 MB of model weights with WebGPU or 26 MB for the CPU path, plus runtime files. Browser caching can make later runs faster.

Removal works best when one person or a clear foreground subject is visually distinct from the background. Fine hair can produce a soft alpha edge, while glass, motion blur, deep shadows, groups, products, and similar foreground/background colours are harder. Preview the cutout on light and dark surfaces before downloading the transparent PNG.

FAQ

Is my image uploaded to a server?

No. The selected image is decoded and processed in a worker inside your browser. The worker downloads the Transformers.js runtime and public MODNet files from external hosts, but your image is not included in those requests. CanDoYa does not receive or store the source or result.

Is this background remover free?

Yes. There is no account, payment, credit system, watermark, or resolution-based export fee. Your device supplies the processing power. The initial public model download still uses your internet connection, which matters on a metered plan.

What image limits apply?

The tool accepts one JPEG, PNG, or WebP image up to 25 MB and 24 megapixels. Those limits reduce browser crashes on memory-constrained phones and laptops. The downloaded PNG keeps the source width and height, although its file size can differ because PNG uses lossless compression.

Does it work without WebGPU?

Yes, on browsers that support WebAssembly and module workers. The worker tests WebGPU first for faster inference. If no usable adapter is available or WebGPU processing fails, it retries with the CPU/WASM backend. That fallback is private too, but can be noticeably slower.

How accurate is the AI background removal?

MODNet is strongest on portraits and clear single subjects. It can preserve soft hair better than hard colour selection, but no automatic matte is perfect. Busy backgrounds, transparent objects, groups, motion blur, shadows, and matching foreground colours can produce missing or retained areas.

Why does the first run download a model?

The neural network has to be available on your device before it can infer an alpha matte. WebGPU uses about 13 MB of FP16 weights; the compatible CPU path uses about 26 MB of FP32 weights, plus AI runtime files. Browsers often cache them, but private mode or storage cleanup can require another download.

Why is the result a PNG?

PNG supports an alpha channel, which stores transparent and partly transparent pixels. Standard JPEG cannot preserve transparency and would fill the removed area with a solid colour. The PNG output also avoids another lossy compression pass on the visible subject.

Can I use the result offline?

After the required runtime and model files are cached, a later run may work without another download, depending on the browser and its storage policy. Do not rely on permanent offline availability: browsers can evict cached model data, and private browsing usually discards it.