Workflows

Workflow: Amazon main image with pure white #FFFFFF

Amazon's main image policy is unambiguous: RGB 255/255/255 background, garment fully visible, no other elements. Here is the Apiway workflow that ships compliant Amazon main images consistently — White Studio + the pure-white post-processing pipeline.

Last updated

Amazon's main image policy for clothing is unambiguous: RGB 255/255/255 background, garment fully visible, no other elements, no logos, no models with text or graphics. This workflow walks the fastest Apiway path to a compliant Amazon main image.

Steps

  1. Open White StudioGo to /app/Creative/ai-photoshoots.open in Apiway
  2. Upload the garmentFlat-lay, mannequin, or on-model input — Apiway will normalize.
  3. Pick aspect ratio 1:1 (Amazon main)Amazon main images are square; pick the 1:1 aspect ratio.
  4. Pick 2K or 4K resolutionAmazon accepts ≥ 1600×1600 px; 2K is more than enough.
  5. Generate and verify the corner pixelAfter generation, sample the corner pixel — Apiway forces #FFFFFF via the post-processing pipeline regardless of the underlying model output.

Amazon image specification (2026)

  • Background: pure white #FFFFFF (RGB 255, 255, 255). Off-white or grey backgrounds get rejected.
  • Aspect ratio: square (1:1) recommended.
  • Minimum size: 1600 px on the longest side for zoom; 2K from Apiway is more than enough.
  • Garment coverage: garment should fill 80–85% of the frame.
  • No props, mannequins (visible), or text overlays.

Why Apiway hits #FFFFFF reliably

Even the strongest underlying AI model (Gemini 2.5 Flash Image / Nano Banana) returns a slightly off-white background — we measured #F2F2F2 to #ECECEC across hundreds of outputs. Apiway runs a separate post-processing pipeline after the model finishes — segmentation, recomposite onto literal #FFFFFF, edge refinement, tone correction. The corner pixel of the final output is true #FFFFFF every time. Read the deep-dive: why we re-composite onto pure white.

Alternative: Ghost Mannequin

For products where the garment is the hero (no model needed), use Ghost Mannequin. Same #FFFFFF guarantee, no model in frame.

Batching for big catalogs

Use Batch Creation when shipping more than ~10 SKUs at a time — runs the same configuration across up to 50 inputs per request.

Related docs