Photo editor for shop owners
Free. Your photos never leave your device. Crop to shop sizes, fix the light, make the background white.
Loading the photo editor. It needs JavaScript and runs only in this browser.
What it does: the handful of jobs a shop owner opens Photoshop for. Crop and straighten to Shopify, Instagram, Etsy and Pinterest sizes, fix light and colour (with a "pick white" dropper and one-click Auto fix), turn a plain backdrop pure white, dab out dust and lint, add a small logo, then apply the same edits to a whole batch and download them as a ZIP.
Your photos never leave your device. Everything happens in this browser. Nothing is uploaded, we cannot see your photos, and there is nothing to sign up for. The photos themselves are not kept after you close the tab. Your edit settings for each photo (not the pixels) are kept in this browser, so if the page reloads you can open the same photos again and the edits come back.
What Photoshop still does that this does not: layers, cutting out hair, generative fill, retouching skin or wrinkles, and colour-managed print work.
How the white background works: a small AI model looks at the photo and finds the product, right here in your browser. The first time you turn it on, your browser downloads the model and the code that runs it from this site (about 16 MB of files, usually about 7 MB as sent compressed) and keeps them for next time. Your photo is never sent anywhere. Only the narrow band along the product's outline is then fine-tuned on the photo's own colours; colour rules never decide large areas, because on pale products that is what used to eat them.
What we measured: an independent check ran it on 38 test photos with known answers (24 built from real garment photos and fabric on shaded backdrops, and 14 drawn product shots). With no taps, 30 of the 36 it worked on lost under 1 percent of the product with no gap over 400 pixels; it refused two tiny shiny items and left them as they were. The misses were light parts of the product on a light backdrop: a whole white sole (a quarter of the shoe), a white panel, a white print bar inside a shirt, white stripe ends, metal buckles and a highlight on a glossy rim. So when a removed area next to the product is lighter than the backdrop around it, the tool marks the photo "Check this one", and a batch ZIP keeps that photo's original background unless you approve it. On those 38 photos plus our own 21 it marked all 7 losses over 1 percent and none of the 48 clean photos; it did not mark the rim highlight (0.8 percent). Those numbers come from the photos the check was tuned on, so look at every result anyway. A Keep tap now grows from where you tap through the photo's own colours: on the sneaker, one tap brought the sole back from 25 percent of the shoe lost to 2 percent, and a second tap to under 1 percent. A Keep tap can also bring back some backdrop next to the part it restores (on the sneaker, a tap in a different spot brought back a white patch above the toe); it shows in the review, and one Undo removes it. The white panel took 2 taps, the print 1, the buckles 3, and the stripe ends 6. What can stay behind: a soft shadow next to the product, backdrop shut in by the product (lace holes, a narrow gap) and tile or floor lines touching it. A tap with Clear an area takes most of that away, and the brush does the rest. If it cannot find a product it can trust, it leaves the photo as it was and says so. Hair and see-through fabric need the brush or a bigger tool.
Credits: the model is U2-Net (small version, u2netp) by Xuebin Qin and colleagues, under the Apache License 2.0, in the ONNX format published by the rembg project (MIT License; notice). It runs on ONNX Runtime Web by Microsoft (MIT License; third-party notices). Both are served from this site. We also have a free invoice maker and a free CRM.
Want your product photos and listings handled for you?
We set up shop listings, product data and the small automations that keep a store tidy, so you spend your time on the products.