Background-remove product photos in bulk for Mercari / EC listings
Tomoda HinataTool author & maintainerPublished Apr 18, 2026Updated Apr 26, 20269 min read
Consistent white-background product photos lift listing conversion 15–30% across marketplaces. Doing 50+ shots in Photoshop is half a day; in-browser AI background removal does it in 10 minutes. The flow: shoot with separation, batch-remove, pad to a fixed canvas with 8–12% margin, keep the alpha master, sharpen + JPG @ 88 for delivery.
Tools used in this guide
Why white background?
Marketplaces (Mercari, Rakuten, Amazon, eBay, Yahoo Shopping) display search results as a thumbnail grid. Photos with consistent white backgrounds visually pop relative to busy in-context shots. Multiple seller-side studies place the conversion lift at 15–30% on listings where the only change was a clean white background.
Shoot for clean edges
Even AI background removal needs clear subject-vs-background separation. Use a plain wall or a large piece of white/grey cloth. Avoid shadows that touch the subject's edges — the AI can be tricked into 'cutting' along the shadow rather than the actual silhouette.
Batch process locally
Our AI background remover handles up to 10 photos concurrently. The model runs entirely in your browser via WebAssembly — no upload, no per-image fee, no rate limits. For a 50-listing run that turns a half-day Photoshop ritual into ~10 minutes.
Standardize size and padding
Marketplaces reward consistency. Pick one canvas (1000×1000 is the broad sweet spot — Mercari accepts square, Amazon requires it) and pad every output with 8–12% whitespace on each side. Our resize tool does this in a single setting.
Keep the alpha master
Export PNG with the alpha channel preserved as your master. Marketplaces that reject transparency get a JPG flattened to white; ones that allow it (and editorial uses) get the original. The reverse — recovering transparency from a flattened JPG — is impossible.
Final pass: sharpen + compress
Marketplace thumbnails downsample aggressively, blurring soft edges. A light sharpen before export keeps text and product silhouettes crisp at small sizes. JPG quality 88 keeps the file lean (250–500 KB typical) without visible artefacts.
Steps
About 10 minShoot with separation
Plain wall or cloth, even front light, no shadows touching the subject's edges.
Batch background-remove
Drop the photos into the AI background remover; up to 10 process concurrently.
Pad to a uniform canvas
Resize all outputs to 1000×1000 with 8–12% whitespace margin.
Save the alpha master
Export PNG with transparency as your archive copy.
Export JPG for upload
Flatten to white, light sharpen, JPG quality 88. Upload.
Frequently asked questions
Will the AI cut hair / fur correctly?
Modern in-browser models handle hair edges surprisingly well — better than 2020-era cloud services for similar inputs. Backlit hair on dark backgrounds is still the failure case; reshoot with even front lighting if the matte fails.
What resolution should I shoot at?
Shoot at the phone's full resolution (typically 12 MP). The AI benefits from extra pixels at the edge transitions. Downsize to 1000×1000 only at the end of the pipeline, not before background removal.
Does the model upload my product photos?
No. The model file (~20 MB) downloads once on first use, then everything runs locally via WebAssembly. Your photos never leave your device.
Mercari requires square photos. What size?
1080×1080 is universally safe. Mercari upscales smaller uploads but cannot recover lost detail; 1080² is the sweet spot.
Should I save as PNG or JPG for marketplaces?
JPG quality 88 for upload (smaller, faster, identical at thumbnail size). PNG with alpha for your master archive.
Try it now
One-click subject extraction
AI Background Remover