Product Photo White Background Maker

📖 What this does

Marketplaces are specific about what a main image may look like: the backdrop has to be pure white, the product has to fill the frame without touching the edges, and any promo message has to sit in a corner. Doing that by hand means a pen-path cut-out, a background fill, a drawn badge, an alignment pass and an export - ten minutes or more per photo, and every one of them comes out slightly different. This moves the whole line into the browser. Drop in a product photo and the background is detected and removed onto white; then stack promo badges into the corners with shape, colour, copy and placement all under your control; then export at the sizes Taobao, Tmall, JD, Amazon and the rest ask for. Everything runs locally, so no photo is uploaded anywhere.

The cut-out is a region-growing pass rather than a neural net. It samples the dominant colour along the four edges, then grows inward from the border, clearing only what is connected to it - so a white label inside a white box survives, and reflective glass or metal keeps its detail. Two extra controls deal with the edge itself: feather softens the transition, and de-fringe eats inward to kill the grey halo that pale products pick up from a light backdrop. Where the algorithm guesses wrong, switch the touch-up brush to erase or restore and paint straight on the preview. Badges stack freely, each with its own shape, fill, outline, type, shadow and opacity, and the move buttons set which one lands on top.

✨ Features

• Auto cut-out onto white - tolerance, edge feather and de-fringe are three separate sliders. Tolerance decides how close a colour has to be to count as backdrop, feather softens the transition, and de-fringe pulls the leftover halo inward.
• Three backdrop modes - pure white, a custom colour, or transparent. Transparent exports as PNG and drops straight onto a detail-page template or a poster.
• Ten badge shapes - rectangle, rounded rectangle, circle, ellipse, triangle, swallowtail, streamer, starburst, shield and speech bubble, plus ten quick templates (new, hot, sale, free shipping, limited time, bestseller, genuine, flash sale, discount, spend and save) applied in one click.
• Every badge is independent - multi-line copy, shape, a nine-point anchor, X and Y offsets, width, height, rotation, solid or gradient fill, outline colour and width, dashed outline, text colour, size, weight and outline, opacity, drop shadow and blur.
• Marketplace canvases built in - Taobao 800, Tmall and Shopee 1000, JD 1200, Amazon 2000, detail page 750x1000 and social 1080, or type your own width and height.
• Live composition feedback - product scale, vertical alignment and both offsets shape the layout, and the status line shows what percentage of the frame the product fills as you go.
• Export, copy or print - PNG, JPG or WebP with a quality setting, plus copy to clipboard and a print button.

🛠 How to use it

1. Press Choose photo or drop a file on the dashed box. With nothing to hand, press Load sample photo to walk the whole flow.
2. Adjust cut-out tolerance against the preview. Raise it if background survives, lower it if part of the product disappears. Raise de-fringe for a grey edge, raise feather if the cut looks too hard.
3. For spots the algorithm cannot judge, switch the touch-up brush to erase background or restore product and drag across the preview; the brush size is adjustable.
4. Pick a marketplace size under Canvas size, then set scale and offsets, watching the fill percentage in the status line.
5. Press Add badge, take one of the quick templates or write your own copy, then tune shape, placement, colour and size. Add more badges as needed and use move up and down to set the stacking order.
6. Choose a format and quality, press Download result, or copy and print instead.

⚠️ Before you rely on it

• The cut-out compares colours, it does not segment by AI. A backdrop close to the product colour, a busy background, or a product with transparent or cut-out sections will all degrade the result - plan on the touch-up brush for those.
• Do not jump de-fringe to a high value. Past a point it eats into the product itself, which shows first on soft edges like plush, fur or feathers. Start at 1 and raise it while watching the preview.
• Very large photos are downsampled first. A thirty or forty megapixel original is reduced to roughly 2.5 megapixels for the cut-out and then composited at the target size, so edge precision is a little below what a small photo gets.
• Use PNG for a transparent backdrop. JPG has no alpha channel and transparent areas come out white.
• Badge sizes are relative. Width, height and text size are all percentages of canvas width, so badges scale when you change canvas size - but check placement again at extreme aspect ratios.
• Everything happens locally. Reloading the page loses the photo and every badge setting, so download anything you care about.
• Two outline settings depend on each other. Outline colour and dashes only appear once outline width is above zero, and shadow blur only applies once drop shadow is ticked - check both switches before concluding something is broken.