How to Create AI Product Listing Images That Convert: Platform-by-Platform Guide (2026)
Wanderson Jackson
Updated July 2026. This guide covers the exact image specs for Amazon, Shopify, Etsy, and DTC stores, then walks through a repeatable AI workflow to produce listing-ready product images at scale. No studio required.
AI-generated product listing images cost 90-97% less than traditional studio photography while meeting the same marketplace specs. The workflow: photograph your product once (even with a phone), describe it precisely in a prompt, generate with a capable image model, then post-process to match platform requirements. offers multiple image models in a single credit pool, so you can test which style converts best for your product category without switching tools.
Why AI Listing Images Beat Traditional Photography on Cost
Traditional product photography for ecommerce averages $50 to $150 per image at a professional studio (source: nightjar.so). For a catalog of 50 SKUs with 7 images each (Amazon recommends 6-7), that is $17,500 to $52,500 before any retouching.
AI image generation flips the economics. A single image generated through a 1-credit model on Avocado AI costs roughly EUR 0.10 to EUR 0.13 on the Starter plan (300 credits for EUR 39/month). For 350 listing images, that is about EUR 35 to EUR 45 total, not per image.
The tradeoff: AI images need human review. Product accuracy matters for marketplace compliance. Amazon requires images to "clearly represent the product" (sellercentral.amazon.com, 2026). AI-generated backgrounds and lifestyle shots are acceptable, but the product itself must look right. This guide covers how to get both the cost savings and the accuracy.
Every marketplace has different specs. Generating images that do not meet platform requirements wastes time and credits. Here is what each platform demands.
Amazon Product Listing Images
Amazon has the strictest main-image rules of any marketplace.
Main image (slot 1):
Pure white background (RGB 255, 255, 255)
Product fills at least 85% of the frame
No text, logos, watermarks, borders, or graphics
No mannequins, hangers, or props
Only the product being sold (no extra accessories unless included)
Minimum 1,000 pixels on the longest side (2,000+ recommended for zoom)
Detail/close-up shots, packaging, and size-comparison shots recommended
No Amazon logos or trademarks
Source: sellercentral.amazon.com/help/hub/reference/external/G1881, verified July 2026.
Shopify Product Images
Shopify is more flexible but has its own best practices.
Recommended size: 2,048 x 2,048 pixels (square, 1:1)
Minimum for zoom: 800 x 800 pixels
Maximum: 5,000 x 5,000 pixels, 20 MB file size
No strict background rule, but white or light backgrounds are standard for product pages
Hero/banner images: 1,920 x 1,080 pixels (16:9)
Source: help.shopify.com/en/manual/products/product-media, photoroom.com/blog/guide-to-shopify-image-sizes, verified July 2026.
Etsy Product Images
Etsy requires images to "accurately represent the item being sold." No strict pixel minimums, but:
Recommended: at least 2,000 pixels on the shortest side for quality
Square or 4:3 aspect ratios work best
White or neutral backgrounds preferred but not required
No watermarks or promotional text
First image is the thumbnail and must show the product clearly
Source: etsy.com/seller-handbook, verified July 2026.
Quick Reference Table
Requirement
Amazon Main
Shopify
Etsy
Aspect ratio
1:1 square
1:1 square
1:1 or 4:3
Min pixels
1,000 (2,000+ recommended)
800 (2,048 recommended)
2,000 shortest side
Background
Pure white (RGB 255,255,255)
White preferred
White/neutral preferred
Max file size
10 MB
20 MB
20 MB
Format
JPEG, PNG, TIFF, GIF
JPEG, PNG, GIF, WebP
JPEG, PNG, GIF
Text overlay on main
No
Allowed
No watermarks
Product fill
85%+ of frame
Best practice
Best practice
The 5-Step AI Product Listing Image Workflow
This workflow produces marketplace-compliant listing images from a single product photo.
Step 1: Capture a Clean Reference Photo
Start with a well-lit phone photo of your product against a simple background. This does not need to be studio quality. What matters:
Even lighting (no harsh shadows across the product)
The full product visible, filling most of the frame
Accurate color representation
No other objects overlapping the product
Natural daylight near a window works well. Shoot from multiple angles: front, 45-degree, top-down. These reference photos anchor the AI generation so the product stays accurate.
Step 2: Write a Precise Product Description
AI image models produce better output when the prompt names specific details. For a listing image, include:
Product type and material (e.g., "matte black ceramic mug with ergonomic handle")
Dimensions or proportions ("12oz capacity, cylindrical body, slightly tapered base")
Color in natural language ("warm terracotta with a subtle speckled glaze")
Surface texture ("brushed stainless steel with fine horizontal grain")
Label or logo placement ("embossed logo on the bottom, no visible branding on sides")
Avoid generic descriptions like "nice product photo." The model will invent details you did not specify.
Step 3: Generate with an AI Image Model
Use a model that handles product subjects well. On Avocado AI, the strongest options for listing images are:
GPT-Image 2 (2 credits): Best text rendering and photorealism. Handles reflections, transparent packaging, and label accuracy well.
Nano Banana 2 (1 credit): Fast, clean detail. Good default for simple products on white backgrounds.
Seedream 5.0 Pro (2 credits): Strong with dense layouts and multi-language text. Useful for products with packaging text in non-English markets.
Recraft V4 (1 credit): Design-focused with style control. Good for lifestyle and styled compositions.
See the model comparison table below for full details.
Step 4: Post-Process for Platform Compliance
After generation, check and fix:
Aspect ratio: Most models output square or near-square. Verify with PIL and crop to 1:1 if needed.
Background purity: For Amazon main images, the background must be pure white. Use the AI model to generate with a white background prompt, or use a background-removal tool as a second pass.
Product fill ratio: The product should fill 85%+ of the frame (Amazon requirement). Crop tightly if the generated image has too much whitespace.
Resolution upscale: If the output is below 2,000 pixels, upscale before uploading. Avocado AI includes AI upscaling on Starter plans and above.
Step 5: Upload and Test
Generate 2-3 variants per product (different backgrounds, angles, or lighting). Upload to your marketplace and track click-through rate. Amazon Brand Registry members can A/B test main images directly. On Shopify and Etsy, rotate variants and compare conversion over 2-week windows.
Choosing the Right AI Image Model for Listings
Not all image models handle product photography equally. Here is how the models available on Avocado AI compare for listing-specific use cases.
Model
Credits
Best For
Text Accuracy
Product Detail
White BG Support
GPT-Image 2
2
Main listing images, packaging with text
High
High
Good (prompt "pure white background")
Nano Banana 2
1
Clean product shots, simple shapes
Medium
High
Excellent
Nano Banana 2 Lite
1
Quick drafts and batch previews
Low
Medium
Good
Ideogram V3
2
Products with logos or on-image text
Very High
Medium
Good
Seedream 5.0 Pro
2
Multi-language packaging, dense layouts
High (14 languages)
High
Good
Recraft V4
1
Styled lifestyle compositions
Medium
High
Good
SeedDream 5.0 Lite
1
Lifestyle product shots
Medium
Medium
Good
All models are available on every Avocado AI plan, from Intro (EUR 19.99/month) through Pro (EUR 249/month). Plans are credit-bounded: Intro gives 100 credits/month, Starter gives 300, Growth gives 800, Pro gives 2,000.
For Amazon main images where the product must be accurate and the background must be white, GPT-Image 2 or Nano Banana 2 are the strongest choices. For lifestyle secondary images, Recraft V4 or Seedream 5.0 Lite offer more creative range.
Writing Listing-Specific Image Prompts
Listing image prompts need more precision than creative or editorial prompts. Here is the pattern that produces consistent, marketplace-ready output.
[Lighting: studio softbox, even illumination, no harsh shadows]
[Background: pure white seamless backdrop]
[Composition: product fills 85% of frame, centered]
[Negative: no text, no logos, no props, no reflections from other objects]
Example: Coffee Mug (Amazon Main Image)
"A 12oz matte black ceramic coffee mug with an ergonomic C-shaped handle, cylindrical body tapering slightly at the base. Shot from a 30-degree front angle with an 85mm lens. Even studio lighting from two softboxes, no harsh shadows. Pure white seamless background. Product centered, filling 85% of the frame. No text, no logos, no props, no watermarks."
Example: Skincare Bottle (Shopify Product Image)
"A 50ml frosted glass serum bottle with a matte gold dropper cap, slightly viscous amber liquid visible through the glass. Shot straight-on from the front at eye level with a 50mm lens. Soft directional light from the left, gentle shadow to the right. Clean white marble surface, white background. Product fills the center of the frame. No text overlays, no brand names rendered as image text."
Example: Leather Wallet (Etsy Listing Image)
"A hand-stitched full-grain brown leather bifold wallet with visible saddle-stitching in cream thread, slightly worn patina suggesting natural aging. Shot from a 45-degree overhead angle with a 35mm lens. Warm natural daylight from above, soft shadows. Off-white linen background. Generous negative space around the product. No text, no logos, no model."
Common Prompting Mistakes
Too vague: "Product photo on white background" gives the model no product information. It will invent a generic object.
Hex color codes: Models interpret natural language better than hex. "Warm terracotta" outperforms "#E2725B."
Brand names in prompt: Do not ask the model to render specific brand logos or text. It will hallucinate garbled text. Use post-processing to overlay real logos if needed.
Asking for Amazon compliance in the prompt: The model does not know Amazon's 85% fill rule. Generate a clean product image, then crop to comply.
Post-Processing for Marketplace Compliance
AI-generated images rarely pass marketplace specs without a quick post-processing step. Here is the checklist.
Background Cleanup
For Amazon main images, the background must be pure white (RGB 255, 255, 255). If the AI model produced a near-white or slightly off-white background:
Use a background remover (Avocado AI includes this via AI Edit)
Replace with a pure white layer
Verify the RGB values at the image edges
Aspect Ratio and Crop
Most AI models output square or near-square images, which works for Amazon and Shopify. For Etsy, 4:3 may perform better in search thumbnails. Crop accordingly.
Crop formula for 1:1 from a non-square output:
python
from PIL import Image
img = Image.open("product.png")
w, h = img.size
side = min(w, h)
left = (w - side) // 2
top = (h - side) // 2
img = img.crop((left, top, left + side, top + side))
img = img.resize((2048, 2048), Image.LANCZOS)
img.save("product_square.png")
Resolution Check
Amazon accepts minimum 1,000px but recommends 2,000+ for zoom. Shopify recommends 2,048px. If your AI output is below these thresholds, upscale with Avocado AI's built-in AI upscaling (available on Starter plan and above) before uploading.
File Naming
Amazon has a specific naming convention: product identifier + variant code + extension (e.g., B08XYZ1234.MAIN.jpg). Shopify and Etsy accept any filename, but descriptive names help with SEO (e.g., black-ceramic-mug-12oz-front.jpg).
Batch Production with Avocado AI Workspace
When you have 20+ SKUs to photograph, generating one image at a time is slow. Here is how to scale.
Using Avocado AI Workspace
The Workspace lets you organize images by project, track generations across models, and manage your credit spend in one place. For batch listing production:
Create a project per product category (e.g., "Kitchen - Mugs," "Skincare - Serums")
Generate 3 variants per product (front shot, 45-degree, lifestyle context)
Review and pick the best variant for each slot
Use AI Edit to fix minor issues (background cleanup, color correction)
Using Flows for Automated Pipelines
Flows lets you chain AI operations. A listing image flow might:
Generate product image with Nano Banana 2 (1 credit)
Apply background removal via AI Edit
Replace with pure white background
Upscale to 2,048px
This runs as a single automated pipeline instead of four manual steps.
Using the MCP Server for Programmatic Access
The Avocado MCP server exposes all image generation tools as programmatic endpoints. If you have a product catalog in a spreadsheet or database, you can script batch generation directly: loop through SKUs, generate images with consistent prompts, and download results.
Credit Planning for Batch Work
SKU Count
Images per SKU
Total Images
Est. Credits (Nano Banana 2)
Est. Credits (GPT-Image 2)
20
7
140
140
280
50
7
350
350
700
100
7
700
700
1,400
500
7
3,500
3,500
7,000
On the Starter plan (300 credits/month, EUR 39), generating 140 images with Nano Banana 2 costs 140 credits, leaving 160 for other work. For 500+ SKUs, the Growth plan (800 credits/month, EUR 99) or Pro plan (2,000 credits/month, EUR 249) makes more sense. Unused credits roll over for one year on all plans.
What Actually Matters
The image is the first thing a shopper sees. On Amazon, the main image determines whether someone clicks. On Shopify, product photos are the primary trust signal. On Etsy, visual quality is the differentiator.
AI image generation does not replace good product photography judgment. It replaces the studio, the photographer, and the retouching bill. You still need to:
Know your product (accurate descriptions lead to accurate images)
Know your platform specs (Amazon's 85% rule, Shopify's 2,048px recommendation)
Test what converts (generate variants, track CTR, iterate)
The cost difference is not marginal. It is 50-150x cheaper per image. For a growing ecommerce brand managing dozens or hundreds of SKUs, that gap funds the rest of the marketing stack.
How to Pick an AI Image Workflow in Under 30 Seconds
Selling on Amazon? Use GPT-Image 2 for main images (best text and detail accuracy). Always verify the white background.
Shopify store with a large catalog? Nano Banana 2 at 1 credit per image keeps costs low. Use Flows to automate the background cleanup pipeline.
Etsy seller with handmade goods? Seedream 5.0 Lite or Recraft V4 for lifestyle compositions that show the product in context.
Products with packaging text? GPT-Image 2 or Ideogram V3 handle on-image text best. Seedream 5.0 Pro supports 14 languages.
Budget-conscious, testing AI for the first time? Start with Nano Banana 2 Lite (1 credit, fastest). Upgrade to Nano Banana 2 or GPT-Image 2 once you know the workflow works.
Need to generate 100+ images per month? Avocado AI Starter (300 credits/month, EUR 39) covers 300 images with 1-credit models.
Running a full creative pipeline (images + video + audio)? Growth plan (EUR 99/month) includes Workspace, Flows, and Storyboards alongside all image models.
Agency managing multiple brands? Pro plan (EUR 249/month, 2,000 credits) plus shared credit pools for team members.
FAQ
Can I use AI-generated images as Amazon product listing images?
Yes. Amazon permits AI-generated images as long as they meet the same requirements as any other product image: pure white background (RGB 255, 255, 255), product fills at least 85% of the frame, no text or logos on the main image, and the image "clearly represents the product" being sold (sellercentral.amazon.com, 2026). AI-generated lifestyle and infographic images are allowed in secondary slots.
What size should AI product images be for Amazon?
Amazon requires a minimum of 1,000 pixels on the longest side, but recommends 2,000 pixels or more for zoom functionality. The main image must be square (1:1 aspect ratio). Generate at the highest resolution your model supports, then resize to 2,048 x 2,048 pixels for optimal display.
How do I make sure AI-generated product images look accurate?
Use a real photo of your product as a reference, and include precise physical details in the prompt (material, color, texture, proportions). GPT-Image 2 and Nano Banana 2 handle product subjects well. Always review the output against the actual product before uploading. Marketplace platforms like Amazon can suppress listings with inaccurate images.
What is the best AI model for ecommerce product images?
It depends on the use case. GPT-Image 2 (2 credits on Avocado AI) is strongest for products with packaging text or complex details. Nano Banana 2 (1 credit) is the best value for clean product-on-white shots. Recraft V4 (1 credit) works well for lifestyle and styled compositions. All are available on every Avocado AI plan.
How much does it cost to generate product listing images with AI?
On Avocado AI, 1-credit models (Nano Banana 2, Recraft V4) cost roughly EUR 0.10 to EUR 0.13 per image on the Starter plan (EUR 39/month for 300 credits). 2-credit models (GPT-Image 2, Seedream 5.0 Pro) cost roughly EUR 0.20 to EUR 0.26 per image. Compared to traditional studio photography at $50-150 per image, AI generation is 50-150x cheaper. Unused credits roll over for one year.
Can I batch-generate product images for my entire catalog?
Yes. Avocado AI's Workspace organizes images by project, and Flows can automate multi-step pipelines (generate, remove background, replace with white, upscale). The MCP server also allows programmatic batch generation for teams with existing product databases. A Starter plan covers 300 images per month; Growth covers 800; Pro covers 2,000.
Do I need to remove the background from AI-generated product images?
For Amazon main images: yes, the background must be pure white. Many AI models can generate with a white background if prompted correctly, but a post-processing cleanup step ensures compliance. For Shopify and Etsy, white or neutral backgrounds are preferred but not strictly required.
What AI image models are available on Avocado AI?
Avocado AI includes GPT-Image 2, Nano Banana 2, Nano Banana 2 Lite, Ideogram V3, Recraft V4, Recraft V4 Pro, Recraft V4 Vector, Seedream 5.0 Pro, Krea 2 Large, Krea 2 Medium, Grok Imagine, SeedDream 5.0 Lite, and Flux 2 Flex. All models are available on all plans. See the full list at avocadoai.co/pricing.
Wanderson Jackson is the founder of Avocado AI, a creative workspace for AI image, video, and audio generation. He writes about ecommerce content production and AI tools for marketers.
