Amazon India AI Image Suppression: Fix Ethnic Wear Listings Fast | Twiink

What triggers Amazon India AI image suppression for ethnic wear?

Amazon India runs two independent suppression systems simultaneously. A listing is auto-suppressed the moment it fails either one — and both apply to every ASIN at the same time.

The first is the Women Ethnic Wear image specification (GE6ZL7C8EDXUPFRL on Seller Central), which governs sarees, kurtas, lehengas, and salwar suits. The spec requires the product to fill 90% of the image area on a pure white (#FFFFFF) background at a minimum of 1,000 px on the longest side. Any single deviation fires automatic ASIN suppression without human review.

The second is Amazon's new global rule effective 23 July 2026 (CNBC, July 2026) requiring photorealistic AI-generated people in product images to carry IPTC 2025.1 metadata before upload. Skip the tag and Amazon's classifier demotes the listing in search and recommendation slots — the listing stays live but loses traffic without any suppression notice.

Step 1: Hero slot — product-only flat-lay or ghost mannequin, nothing else

The hero image (position 1) is the strictest checkpoint on any Amazon ethnic-wear ASIN. The rule is simple: no model, AI or real, in the hero slot. The image must show only the product — a flat-lay or ghost-mannequin shot of the saree, kurta, or lehenga on pure white.

The six most common hero-slot suppression triggers, per Amazon's consolidated suppressed-listing guide:

• Non-pure-white background — off-white (#FAFAFA), cream, grey, or any gradient. Must be exactly RGB #FFFFFF.
• Product fill below 90% — the garment must occupy at least 90% of the frame for ethnic wear. Excessive negative space triggers suppression.
• Resolution below 1,000 px — Amazon's zoom feature requires 1,000 px minimum on the longest side; 2,000-2,400 px is recommended.
• Visible model or mannequin — any human figure, AI or real, disqualifies the hero slot for ethnic wear.
• Watermark, text overlay, or logo — any graphic element on the image triggers automatic rejection.
• Packaging or swing tags in shot — remove all tags and packaging before capture.

For a deeper look at when to choose ghost mannequin versus on-model, see ghost mannequin vs AI model for ethnic wear India — the hero slot is always ghost-mannequin territory.

Step 2: Export settings that clear the background and resolution checks

Export format matters as much as the image itself. Use these settings for every Amazon India hero image:

The "1,000 DPI" figure in older Amazon India help pages refers to pixel count on the longest side, not print resolution. The export setting that matters is pixel dimensions — save at 72 dpi metadata.

For sellers managing a saree or kurta catalog across dozens of SKUs, AI catalog photos for ethnic wear with background-removal built into the workflow clear both the background check and the resolution floor in a single export pass.

Step 3: Tag AI-generated on-model images before upload (July 2026 rule)

The 23 July 2026 rule applies only to images containing a photorealistic AI-generated human figure. Flat-lays, ghost-mannequin shots, and fabric swatches are not in scope and need no tagging — only images where a human model was generated by AI require the metadata fields.

For AI on-model images in secondary slots (positions 2-7), embed these four IPTC 2025.1 XMP fields before upload — via Photoshop's File Info panel or the ExifTool command-line utility:

• AISystemUsed — the AI tool that generated the image (e.g. "Twiink AI Virtual Try-On")
• AISystemVersionUsed — the version or build date of the tool used
• AIPromptInformation — the text prompt used to generate the image
• AICreator — the person or brand who wrote the prompt

Amazon's Seller Central listing editor added a dedicated "AI-generated person image" toggle on 22 July 2026 that writes the corresponding tag server-side. You must set both the local XMP embed and the listing toggle — either alone is insufficient to satisfy the rule.

Critical export note: using Photoshop's Save for Web (Legacy) strips the C2PA 2.0 manifest that AI generation tools embed by default, producing an untagged image that triggers demotion even when the IPTC fields were correctly set upstream. The correct path is File > Save As, followed by manual verification that all four IPTC 2025.1 fields are populated.

Step 4: Use secondary slots (2-7) for saree drape, kurta styling, and on-model shots

Amazon ethnic-wear secondary image slots explicitly allow full-body or 3/4 on-model shots — real or AI-generated — as long as IPTC tags are present and the listing toggle is set. This is the compliant home for lifestyle imagery: saree drape detail, kurta layering, dupatta styling, and embroidery close-ups.

The compliant split for a complete ethnic-wear ASIN:

• Position 1 (hero): ghost-mannequin or flat-lay product shot, pure #FFFFFF background, 90% frame fill, 2,000+ px. No model of any kind.
• Positions 2-4: AI on-model shots (virtual try-on) showing front, 3/4, and side drape. IPTC 2025.1 tags required; listing toggle set.
• Positions 5-7: fabric and embroidery close-ups, size-chart infographic, styling detail. No tagging required unless a human figure appears.

For men's ethnic wear catalogs, the same split applies. See AI model photos for men's ethnic wear India for the kurta and sherwani-specific approach to secondary slot imagery.

How long does reinstatement take after fixing the images?

Reinstatement timelines depend on the suppression type. Three distinct outcomes:

• Image quality failure (wrong background, framing, watermark, resolution): approximately 48 hours after a corrected image is uploaded, per Amazon's seller communications.
• AI-image traffic demotion (missing IPTC tag or listing toggle not set): typically 24-72 hours after re-uploading with the AI-person flag correctly set in both the XMP metadata and the listing editor.
• Account-level suspension from repeated or severe violations: first appeal response in roughly 5 business days; full reinstatement in 7-14 days. A separate and significantly more serious outcome than a single image quality flag.

Preventing suppression is faster than recovering from it. For sellers focused on long-term conversion, compliance-first images also reduce return rates by showing accurate drape, color, and fit — see how catalog images reduce ethnic wear returns in India.

How Twiink generates Amazon-compliant ethnic wear catalog images

Twiink's Product Shots feature generates ghost-mannequin and flat-lay images built for Amazon hero-slot compliance by design: single product, no model visible, with Background Replacement setting the output to pure white #FFFFFF. The two most common suppression triggers — non-white background and model in main image — are cleared in one step.

For secondary slots, Virtual Try-On generates full-body and 3/4 on-model ethnic wear images — saree drape, kurta styling, dupatta layering — that are compliant in positions 2-7 once the IPTC tagging workflow is applied before upload.

A studio sample shoot for a saree or kurta catalog at scale runs 3-7 days from booking to final edited delivery. Twiink turns one flat garment photo into hero-ready and secondary-slot images in minutes. For a full breakdown of what studio rates cost at catalog scale, see fashion catalog photoshoot costs in India — the per-SKU comparison makes the case for AI output clearly.