Drop 30 product shots · get 150 platform-ready cuts

Background removal, multi-platform sizing, optional watermark and shadow. The Agent processes the batch in an isolated sandbox and writes the outputs to your Drive in a per-SKU folder tree — Amazon main / variants, Tmall A+ banners, Shopee thumbs, Instagram crops, all named to the platform spec.

A peek at what you get

Before / after
Page 1 of 4·Before / after
Drive folder tree
Page 2 of 4·Drive folder tree
Platform spec sheet
Page 3 of 4·Platform spec sheet
Batch report
Page 4 of 4·Batch report

30 raw shots · 5 platforms · 150 platform-ready cuts

Each platform has its own size, fill ratio, and background rule. The sandbox enforces them — the upload reviewer never sees a violation.

Raw shot
IMG_4821.HEIC
Cut + 3 brand backgrounds
amazon-2000²
ig-square
hero-banner
rembg.process30 images · avg mask conf 0.85
vb image generate3 lifestyle backgrounds · category-matched
run_pythonPillow · 5 platform sizes · sRGB · brand kit

Drive · per-SKU folder tree

Output mirrors what Amazon Seller Central, Tmall Merchant Center, and Shopify product import expect. Hand the tree to your VA — they upload without renaming a single file.

Drive·/product-photos/march-2026/
/product-photos/march-2026/
sku-a41-ceramic-mug-blue/
amazon/
main-2000x2000.pngspec ✓
variant-3q-left.pngspec ✓
variant-3q-right.pngspec ✓
variant-top-down.pngspec ✓
tmall/
main-1500x1500.pngspec ✓
aplus-banner-970x600.pngspec ✓
social/
ig-1x1.png
ig-4x5.png
reels-9x16.png
pinterest-2x3.png
sku-a42-...
... 28 more SKUs ...
_batch-report.pdf2 flagged · 1 retry
_batch-report.pdf
4 pages · 2 flagged · 1 retry succeeded

How it works

Step 01

Drop the raw photos

A zip with multiple SKU folders works best. JPG, PNG, HEIC, WebP — the Agent ingests them as-is. If a photo is misnamed or rotated wrong, it gets flagged in the batch report instead of silently slipping through.

Step 02

Pick the delivery target

Amazon + Tmall main + variants? A full social-media pack? Lifestyle scenes for an IG campaign? Each preset has its own size / background / shadow rules — the Agent enforces them so platform reviewers do not bounce your upload.

Step 03

Pick up the folder tree

The output lands in your Drive at /product-photos/{batch}/{sku}/{platform}/{file}. A batch report sits at the top — flagged photos, skipped SKUs, total cuts per platform, and total processing time per SKU.

Why Vecbase for this

Background removal is a sandboxed pipeline, not a vendor API call

Most batch tools wrap one background-removal API. Here the Agent runs rembg locally in the sandbox, can swap to a generative model when the silhouette is hard (glass, fur, jewelry), and lets you say in chat "redo SKU-A41 with the model that handles transparency better". You watch the swap happen in the transcript.

Per-SKU folder tree means downstream tools just work

The output mirrors what Amazon Seller Central, Tmall Merchant Center, and Shopify product import CSVs expect. Hand the Drive tree to your VA — they upload without renaming a single file.

Brand colors stay consistent across batches

Once you mention the brand backgrounds once ("warm beige is #F5EFE0"), the Agent saves it to your workspace and every future batch uses the same hexes. No more "the new VA used the wrong beige" surprise during a campaign launch.

Batch report calls out what looks off

If a product cutout has a low-confidence mask (chewed-up silhouette), or a photo is rotated wrong, or a SKU folder has fewer angles than its peers, the batch report flags it. You don't discover it three days later from a customer complaint.

Frequently asked

Yes. Tell the Agent in chat once ("our beige is #F5EFE0, shadow is 12px soft, 0.4 opacity, offset 0/6") and it saves that to your workspace. Every future batch reuses the same values automatically. If you switch brand identity, change it once and the next batch flips accordingly.

Get yours in under 90 seconds

Sign in, hand it over to the Agent — the finished file lands in your Drive.