The Export Settings Are the Requirements
Most “product photo background remover” tools stop at the cutout and leave you to find out — usually via a suppressed listing — that their “white” is a soft 245-gray, their canvas keeps the original 4:3 ratio, or their ZIP renames everything to image-1.png. A remover built for ecommerce treats the platform rules as export settings instead: explicit RGB(255,255,255) compositing (not “leave the pixels near-white”), square-crop-to-product-bounding-box so the subject hits the 85% fill bar, sRGB JPEG at your target size, and filename passthrough so SKU-named files stay SKU-named. One pass, one ZIP, no re-touch in Photoshop afterwards.
Cross-Platform Spec Table (verified 2026)
| Platform | Background | Size / ratio | The catch |
|---|---|---|---|
| Amazon (main image) | Pure white RGB 255,255,255 | 1000 px min, 1600+ for zoom, 1:1 preferred | Corner sampling; text/logos/insets cause suppression |
| Walmart Marketplace | Seamless white 255,255,255 | 2200×2200 recommended, 1500 minimum, 1:1 required | Non-square is auto-unpublished, not padded — stricter than Amazon |
| eBay | White or light neutral, no official RGB | 500 px min, 1600 for zoom | No published spec — quality-score ranking rewards consistency |
| Etsy | No rule; white or contextual both fine | 2700×2025 recommended (4:3), 1 MB cap | First image becomes the search thumbnail — square gets letterboxed |
| Shopify | Anything your theme allows | 2048×2048 sweet spot, 20 MB / 25 MP ceiling | Theme crops to square in most galleries — see our banner-size guide |
| Google Shopping | White allowed, non-white OK for many categories | 250×250 to 5000×5000 | image_link must be directly fetchable; redirects get dropped |
Read across the table and one export wins everywhere: a transparent PNG master flattened onto exact white, squared at 2000–2200 px. It clears Amazon and Walmart simultaneously (the only two platforms with hard white rules — see the Amazon white-background breakdownand the Walmart requirements guide for each one's failure modes), and every looser platform accepts it without cropping surprises.
One Master, Many Deliverables
The delivery structure that keeps a catalog maintainable is asymmetric: the master keeps the alpha channel (transparent PNG — the format that guarantees the cutout survives, explained in our transparent format guide), and everything the platforms see is a flattened copy exported from it. White JPEG for Amazon and Walmart, light-gray for a branded Shopify grid, composited scene for Etsy — all from the same mask, zero re-cutting. When a platform changes its mind, you re-export; you do not re-shoot. The bulk background changer page covers the per-platform color values in detail.
Two QC habits make the whole pipeline safe. First, the fake-transparency test: paste any delivered PNG onto a black canvas — if the “background” shows up as a white box or gray fringe, the file was flattened by an earlier tool and is not a master. Second, the corner-pixel test for white deliverables: sample all four corners in any viewer that shows RGB values; all four must read 255,255,255, not 254. BatchBG's white mode composites explicit 255 values rather than relying on the source looking white, which is the step between “why was this listing rejected” and a passing upload.
Shoot for the Cutout, Not Just the Photo
Background removal quality is mostly decided before upload. Four shooting rules pay for themselves on a 300-SKU catalog: leave 10%+ margin around the product (the matting model needs context pixels), avoid a white product on a white surface (low-contrast edges are where every model eats detail — shoot on gray and replace), diffuse the lighting so no hard shadow anchor survives under the product, and shoot above 2000 px so the 85% fill crop still lands on the 1600 px zoom threshold. If you get these wrong, no background remover — local or server-side, free or $2/image — can invent separation that was never in the photo. The step-by-step version of this workflow, from shoot to upload, lives on the product photos page; for whole-folder throughput see the bulk background remover.
Honest Limits
- The hardest 5% of edges. Flyaway hair on a model shot, thin gold chains, and multi-faceted glass can still need per-image touch-up anywhere at this price. HD mode (BiRefNet) narrows that set; it does not empty it.
- Giant folders need RAM. Everything runs on your machine, so a 500-file drop of camera-native RAW-sized JPEGs will thrash a 4 GB laptop. Batch 20–40 files or resize first with the batch compressor.
- No AI scene generation. This outputs a clean cutout or a flat color — if you need a photorealistic lifestyle scene generated for you, that is a different (and much more expensive) tool, and we will say so instead of pretending.