Background Remover
Lift a subject off its background in the browser, with tolerance and feathering you control.
What This Does, And What It Does Not
Most background removers send your photograph to a server running a segmentation model. That works well on anything the model was trained for, and it means the picture — someone's face, a product not yet announced, a passport scan — now exists on somebody else's hardware. This one never sends anything anywhere.
What runs here is a flood fill from the edges: it samples the corners, spreads inward through everything close enough in colour, and stops at the first boundary it cannot cross. On the pictures that matter most for this — a product on a white sweep, a logo on a flat field, a screenshot with a solid backdrop — it is clean, fast and exact. On a portrait against a busy street it is not, and it will not pretend otherwise.
Key features
- Nothing is uploaded — the file is read into the page and never sent.
- Tolerance you control — widen it to catch a gradient backdrop, narrow it to protect a subject that shares the background's colour.
- Feathered edges — a soft alpha ramp instead of the jagged staircase a hard threshold leaves.
- Pick the colour yourself — sample the corners automatically, or click the exact shade to remove.
- Transparent PNG out — ready to drop onto any background, with the alpha channel intact.
This tool runs entirely inside your browser using native Web APIs. Your files and text are never uploaded to a server, never logged and never shared with third parties.
How to use: Background Remover
- Open your image
Drag it in or click to choose. PNG, JPEG and WebP all work. - Let it sample the corners
For a photo on a plain backdrop, the first result is usually close. - Adjust tolerance
Too much background left means raise it; a subject losing its edges means lower it. - Feather, then download
A few pixels of feathering hides the cut. Save as PNG to keep the transparency.
Technical specifications
| Processing location | Entirely in your browser — no server round trip |
|---|---|
| Data uploaded | None. Files and text never leave your device |
| Price | Free — no account, no trial, no usage cap |
| Category | Media |
| Works offline | Yes, once the page has loaded |
| Browser support | Chrome 90+, Edge 90+, Firefox 90+, Safari 15+ |
| Interface languages | English, 中文, हिन्दी, Español, العربية |
Frequently asked questions
Why does it struggle with hair and fur?
Because a flood fill decides per pixel, and hair is thousands of pixels that are partly background. Getting that right needs a trained matting model, which needs either a large download or a server. This tool is honest about the trade: it is excellent on flat backdrops and poor on fine detail, and you will see which you have within a second of opening the file.
The subject has bits of it removed. What now?
Lower the tolerance. That happens when part of the subject is close in colour to the background, and a lower tolerance stops the fill crossing into it. If they are genuinely the same colour, no threshold will separate them and the picture needs to be cut by hand.
Why is the download a PNG and not a JPEG?
JPEG has no alpha channel — it cannot store transparency at all. Saving a cut-out as JPEG would fill the removed area with a solid colour, which defeats the purpose. PNG and WebP both keep the alpha, and PNG is the one that opens everywhere.
Does it work on a screenshot?
Very well, usually. Interface backgrounds are flat by design, so the fill finds a clean boundary. It is one of the cases where a simple algorithm beats a model, because there is nothing to guess at.
Is there a size limit?
Only your device's memory. The image is held as raw pixels while it works, so a 50-megapixel photograph needs a few hundred megabytes. On a phone, resizing first is kinder — and for a cut-out destined for the web you were going to resize anyway.