Last month I listed a 240-SKU product drop on Amazon. The photos were a mess: mixed white balance, random filenames like IMG_4471.JPG, and half still had cluttered backgrounds from my desk.
The standard advice is to use a cloud background-removal service. My problem: at 240 images, per-image credits and upload limits turn a 40-minute job into a $20+ afternoon of babysitting a browser tab.
So I built a batch product photo pipeline that runs entirely offline. 240 images, start to finish, in under 40 minutes. Here is the exact workflow.
Why batch processing beats one-at-a-time tools
The bottleneck in e-commerce photography is never the edit — it is the coordination. Doing 240 images one at a time means 240 uploads, 240 downloads, 240 renames, and 240 chances to fat-finger a filename.
A batch pipeline collapses that into five steps that run while you make coffee.
The 5-step batch workflow
1. Sort and quarantine
Dump every RAW/JPG into one folder, then split by SKU family. I keep a _rejects folder for anything out of focus — you do not want a blurry shot sneaking into a batch of 240.
2. Batch background removal
This is the step everyone pays for. Running it locally removes the per-image cost entirely. I drop the whole folder in, hit run, and come back to transparent PNGs.
Key spec: images stay 1000x1000 px minimum for Amazon zoom, so do not let a tool downscale on the way out.
3. Resize to marketplace specs
Amazon, eBay and Etsy want different things:
- Amazon: 2000x2000 px recommended (enables zoom), pure white background
- eBay: 1600 px on the longest side, 500 px minimum
- Etsy: 2000x2000 px, 2700 px recommended for the primary image
Set these once as presets and the batch applies them consistently. Inconsistency between listing images is the thing that makes a store look amateur.
4. Auto-rename by SKU
This is the quiet hero. Every file gets {SKU}-{index}.jpg. When a customer emails about SKU B07XYZ, you find the photo instantly instead of scrolling a grid of thumbnails.
5. Convert and compress for the web
Export JPG at quality 82 — visually identical at listing size, and it keeps your page weight down. Nothing kills a product page like a 4 MB hero image.
The economics
| Item | Cloud service | Offline batch |
|---|---|---|
| 240 images | $20-40 in credits | $0 per image |
| Upload time | 240 round trips | none |
| Privacy | images leave your machine | stays local |
| Repeat monthly | cost repeats | cost is sunk |
If you list products more than twice a year, a one-time offline tool pays for itself on the first batch.
What I used
I ran this with BulkPhoto AI — a $29 one-time batch photo tool (background removal, marketplace resize presets, SKU renaming) that runs offline. No subscription, no credits.
The full step-by-step write-up with screenshots is here: the complete batch product photo workflow.
Takeaways for developers
- If a task has a per-unit cost and you do it repeatedly, batch it locally.
- Deterministic naming (
SKU-index) is worth more than any AI feature. - Marketplace specs are an API contract — encode them as presets, not as memory.
Originally published on aixhdd.com.
Top comments (0)