Make a product photo background white
Drop in a photo shot against a flat background — studio white, grey, black, or a seamless backdrop — and this tool removes that background and replaces it with pure #ffffff. It runs entirely in your browser, so the photo never leaves your device.
What this tool is good at — and what it is not
This is a flat background remover, not a general-purpose AI cutout. It works by measuring how far each pixel sits from the background colour, so it is extremely fast, completely predictable, and needs no model download — but that also defines its limits honestly.
| Situation | Result |
|---|---|
| Studio shot on seamless white, grey or black | Excellent. Cleaner edges than AI cutouts, because nothing is being guessed |
| Product photographed on a coloured paper sweep | Very good — set tolerance to taste |
| Product on a table with a plain wall behind | Often good, if lighting is even |
| Busy or textured background (room, outdoor, patterned cloth) | Will not work. No tolerance setting fixes this — it needs a segmentation model |
| Subject with soft hair, fur or glass edges | Partial. Increase edge softness; expect to touch up |
| Transparent or reflective products | Weak. There is no background to distinguish from the object itself |
Most catalogue photography falls in the first three rows, which is exactly what this tool is built for. For rows four and five you need a segmentation model, and those come with their own trade-offs — see the note at the bottom of this page.
How to get a clean result
- Start with even lighting. The single biggest factor is not the tool, it is whether the backdrop is a consistent colour. A visible gradient across a white sweep defeats any colour-based method.
- Let the tool detect the background first, then raise tolerance gradually until the backdrop disappears. Stop as soon as it does. Pushing tolerance higher than necessary starts eating into pale parts of the product.
- Keep "Only remove background connected to the image edge" switched on unless you have a reason not to. It is what stops a hole or a dark logo in the middle of a product from being deleted along with the backdrop.
- Leave edge cleanup on. Without it, a product shot on a dark backdrop keeps a grey halo after the swap to white, because edge pixels are a blend of product and backdrop.
- Match the output to the platform. For most marketplaces the answer is JPEG on pure white. Use PNG with transparency only if you intend to composite the product onto something else later.
- Keep the original. This tool exports a new file; the source is untouched. Catalogue work always ends up needing the original again.
The two settings that decide quality
Tolerance
Tolerance sets how different a pixel has to be from the detected background before it is treated as part of the product. Inside the lower threshold a pixel is fully removed; beyond the upper threshold it is fully kept; in between it becomes semi-transparent, which is what gives a natural edge instead of a stair-stepped one.
Too low and background remains as speckle. Too high and light or neutral parts of the product — white labels, chrome, pale packaging — start becoming transparent. There is no universally correct value; it depends on how uniform the backdrop is and how noisy the photo is.
Edge cleanup
This is the step most simple background removers skip, and it is why their output looks wrong when the background was dark. At the boundary, a pixel is physically a mixture of the product and the background. Removing the background does not remove that mixture — so if you photographed on black and composite onto white, the rim of the product keeps a grey fringe.
Edge cleanup solves this by un-mixing the pixel: given the estimated transparency and the old background colour, it reconstructs what the pixel's colour would have been against a neutral backdrop, then composites it onto the new background. The maths is simple, but it has to be applied only where a pixel is genuinely semi-transparent — applied everywhere it would damage texture in fully opaque areas.
How marketplaces judge a white background
Most product photo specifications do not ask for "a white background" in the abstract; they ask for a specific value, and they check it by sampling the pixels around the product.
| What is checked | What passes |
|---|---|
| Background pixel value | Pure white, #ffffff, not "off-white" from a warm light or an uncalibrated monitor |
| Uniformity | No visible gradient, vignetting or shadow cast onto the backdrop |
| Product in frame | Roughly 85% of the frame, centred |
| Edges | No halo in the old background colour, no hard stair-stepping, no leftover speckle |
| Text and logos | None on a main image; allowed on secondary images |
A background that looks white on screen can sample as #f8f6f2 — that is a fail, and it is one of the most common reasons a listing is sent back. Exporting from this tool with the white option guarantees the value rather than hoping the photo got it right.
Composing a full listing image
Removing the background is normally one step of two. A marketplace usually wants a fixed pixel size as well, and once the background is white you may want the product larger in the frame. The usual order is:
- Remove the background here and export as PNG with transparency.
- Resize to the pixel size the platform asks for, padding onto a square white canvas at about 85% fill — the resize tool does exactly this and accepts PNG input.
Doing it in that order matters: resizing a JPEG that already has a white background bakes any halo into the file and can introduce a faint seam at the edge of the padding. Keeping transparency until the last step avoids both.
Questions
Is my photo uploaded anywhere?
No. The pixels are read and rewritten in your browser using the canvas API, and the file is never transmitted. Disconnect your network after the page loads and the tool continues to work — that is the test.
Why can it not handle a complicated background?
Because it decides what to remove by colour, not by understanding what is in the picture. If the backdrop varies from one corner to another, no single background colour describes it, and any threshold that removes the backdrop will also remove matching parts of the product. Handling that requires a segmentation model that has learned what a product is.
Why is there a grey fringe left on my product?
Turn on edge cleanup. A rim of grey is the signature of a dark original backdrop: those edge pixels really are darker than the product, because they are a mix of the two. Edge cleanup reconstructs the underlying colour; without it, you are compositing a mixture.
Can I get a transparent PNG instead of a white background?
Yes — choose "Transparent" and PNG as the output format. Transparency is lost if you export as JPEG, so the two settings need to agree.
Does it work with HEIC files from an iPhone?
Only if your browser can decode them. Chrome and Edge cannot read HEIC directly; Safari on macOS and iOS usually can. If the file does not load, convert it to JPEG first.
Why do you not just use an AI model?
Because of what it would cost you. The accurate open background-removal models, such as RMBG from BRIA, are licensed for non-commercial use only — commercial use requires a paid agreement, and this site is funded by advertising. The alternative is an API, but that means uploading your product photos to somebody else's server, which contradicts the reason this site exists. So this tool solves the case that colour analysis genuinely solves well, and says plainly where it stops.