All guides

How to remove a background from a photo

Every other background remover asks you to upload your photo to a server first. Emaki Studio does the cut on your own device: the subject is detected in the browser, the background is cleared, and you get a transparent PNG. That matters when the picture is a product shot you were trusted with, or your own face. Here is how.

  1. 1. Open Remove Background and add your photo Drop in a JPG, PNG, HEIC or WebP file. The first cut downloads the small on-device model, which takes a moment; after that it is cached and every photo is quick.
  2. 2. Let it find the subject The cut runs automatically - that is what the tool is for, so there is no second button to press. A checkerboard appears behind the result, which is how you can tell transparency from a white background.
  3. 3. Tighten or soften the edge Hair, fur and fine detail are where a cutout usually falls apart. The edge and softness controls re-solve the outline against the photo itself, so strands survive instead of being shaved into a smooth blob.
  4. 4. Export a transparent PNG Download the result as a PNG with a real alpha channel. Save it as JPG and the transparency is lost - JPG has no alpha - so PNG is the format to keep.
  5. 5. Or cut a photo out inside a design The same cutout works in the carousel editor: select a photo, open Properties and remove its background there, then drop it over a colour or a pattern. The original photo is kept, so you can undo the cut at any time.

Open Remove Background · Read the full guide

Is my photo uploaded to remove the background?
No. The model runs in your browser on your own device, so the photo never leaves it. There is no account, no queue and no per-image limit.
Does it work on hair and fur?
Yes - the edge is refined against the photo itself rather than only the detected outline, which is what keeps individual strands. Very fine, low-contrast hair against a busy background is still the hardest case for any tool.
What format should I save a cutout in?
PNG. It carries the transparency; JPG does not and will fill the cleared area with white. WebP also supports transparency if you need a smaller file.
Does it use generative AI?
No. The model classifies which pixels belong to the subject - it never invents pixels. Nothing in your photo is generated or sent anywhere.