How to Use GPT Image 2 for E-commerce

EvoLink Quick Start

Turn an ecommerce prompt recipe into a GPT Image 2 API call:

Model Page · Docs · API Key · GPT Image Prompts · Skill

export EVOLINK_API_KEY="your_key_here"

curl --request POST \
  --url https://api.evolink.ai/v1/images/generations \
  --header "Authorization: Bearer ${EVOLINK_API_KEY}" \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "gpt-image-2",
    "prompt": "A premium marketplace main image for a stainless steel water bottle on a clean white background, realistic reflections, sharp product edges, commercial lighting"
  }'

How to Use GPT Image 2 for E-commerce

A prompt cookbook for creating e-commerce images with GPT Image 2 — listing photos, model try-on, product showcases, interaction scenes, and social commerce creatives.

🍌 Introduction

This repository shows you how to use GPT Image 2 to create e-commerce images — listing photos, multi-country main images, A+ detail visuals, model try-on, product showcases, interaction scenes, and social commerce creatives — at production scale.

Every case below is a recipe: an input product photo, the exact prompt to paste into GPT Image 2, and real output examples. Replace the {placeholders} with your product, language, country, and selling points.

Need to run hundreds of SKUs at once? Jump to 🚀 Batch Generation via Evolink API.

Try it on Evolink: GPT Image 2 workflow

If you find it useful, please give us a Star ⭐

Note

This repository focuses on reusable prompt templates for ecommerce creatives. Replace the placeholders inside curly braces before use, and adjust to fit your actual platform, market, language, product, material, model type, and selling points.

How to Use These Prompts

Pick a case below that matches your e-commerce goal (listing main image, model try-on, multi-angle product, A+ detail visuals, etc.).
Upload your product photo and copy the prompt from that case — replace {placeholders} with your product, language, country, platform, and selling points.
Run it on GPT Image 2 — either through the Evolink workflow UI for one-off creatives, or via the Evolink API for bulk SKUs and multi-language rollouts.

🖼️ Ecommerce Listing Image Prompts

Case 1: Marketplace Listing Image Set

Prompt:

Generate a {platform} listing image set for {country} in {language}. The product is {product_name}, and the selling points are {selling_points}.

Note

Replace {platform}, {country}, {language}, {product_name}, and {selling_points} before use. This template fits Amazon, eBay, AliExpress, TEMU, SHEIN, TikTok Shop, Shopee, Lazada, Mercado Libre, Walmart, Etsy, Rakuten, Coupang, Shopify, and similar marketplaces. The agent will plan platform-appropriate listing assets such as clean white-background images, size diagrams, multi-angle views, lifestyle scenes, selling-point images, and material details.

Example 1: Solid Wood Dining Chair

Input	Output
_{Product photo}	_{Clean background main}	_{Selling points}	_{Size diagram}	_Multi-angle
_{Product photo}	_{Lifestyle scene}	_{Usage scene}	_{Material detail}

Example 2: Sport Water Bottle

Input	Output
_{Product photo}	_{Clean background main}	_Multi-angle	_{Selling points}
_{Product photo}	_{Size diagram}	_{Usage scene}	_{Material detail}

Example 3: Sport T-shirt

Input	Output
_{Product photo}	_{Selling points}	_{Product specs}	_{Size chart}	_Multi-angle
_{Product photo}	_{Usage scene}	_{Material detail}	_{Selling-point summary}

Case 2: Multi-Country Listing Image Set

Prompt:

Generate a {platform} listing image set for {country} in {language}.

Note

Replace {platform}, {country}, and {language} before use. Use this shorter pattern when you want the model to infer the best listing layout automatically.

Example 1: Purple Polka-Dot Dress

Input	Output
_{Product photo}	_Arabic	_Chinese	_English
_{Product photo}	_Japanese	_Korean	_Portuguese

Example 2: Cordless Vacuum

Input	Output
_{Product photo}	_Bengali	_German	_Spanish
_{Product photo}	_French	_Hindi	_Russian

Example 3: Christmas-Theme Phone Case

Input	Output
_{Product photo}	_Indonesian	_Italian	_Dutch
_{Product photo}	_Thai	_Turkish	_Vietnamese

Case 3: A+ Product Detail Visuals

Prompt:

This is a {product_description}. Generate A+ detail visuals for {country} in {language}.

Note

Replace {product_description}, {country}, and {language} before use. Example: "This is a box of coffee beans. Generate A+ detail visuals for the US in English." Useful for product storytelling, feature breakdowns, usage scenes, and premium detail-page layouts.

Example 1: Premium Tea Gift Box

Input	Output
_{Product photo}	_{Hero (selling points)}	_{Design & craft}	_{Product line}
_{Product photo}	_{Usage guide}	_{Gifting scene}	_{Usage scene}

Example 2: Natural Glue Stick

Input	Output
_{Product photo}	_{A+ hero}	_{A+ detail 1}	_{A+ detail 2}
_{Product photo}	_{A+ detail 3}	_{A+ detail 4}	_{A+ summary}

Example 3: Bubble Tea Drink

Input	Output
_{Product photo}	_{A+ hero}	_{A+ detail 1}	_{A+ detail 2}
_{Product photo}	_{A+ detail 3}	_{A+ detail 4}	_{A+ summary}

Example 4: Sony A7 Exploded View (Community, by @iamaiistudio)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet only shared the output image.

Observed prompt:

Detailed exploded-view diagram of a Sony A7 mirrorless camera, with all internal components separated and clearly visible, each part labeled with its name. Technical product illustration style, clean white background, precise and informative layout.

Output
_{Exploded-view product diagram}

Example 5: Architectural Product Catalog Page (Community, by @iamaiistudio)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet shared the prompt and output image.

Observed prompt:

Create a vertical 3:4 product design catalog page with a warm neutral paper-like background.

Top section - lifestyle hero shot: place the product (use the uploaded image as the exact reference, preserving its form, proportions, materials, and identity without redesign) center-dominant with generous whitespace. Setting is a minimal architectural interior with a textured plaster wall and subtle concrete/stone floor. Lighting is natural sunlight angled from the side, soft but casting high-contrast shadows. Render in editorial lifestyle photography style, high realism, warm and muted color grading.

Bottom section - technical specification panel laid out in a clean modular grid:
- Bottom left and center: orthographic architectural line drawings showing front view, side view, and three-quarter cutaway/profile view. Lines in muted red or sepia, fine technical weight, with minimal editorial measurement and construction callouts.
- Bottom right: 3-4 material swatch samples derived from the product's actual materials (fabric, leather, metal, wood, or plastic as applicable), in square or rectangular format with small editorial captions.

Typography: minimal editorial style, subtle captions only, no large headlines, soft black or dark brown.

Overall mood: design catalog / product design journal - architectural, premium, calm. No clutter, no bold colors, no heavy branding, no decorative graphics, no perspective distortion in the technical drawings.

Output
_{Catalog page with technical panel}

Example 6: Reverse Disintegration Product Ad (Community, by @iamaiistudio)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet shared the prompt and multiple output variations; a representative watch composition is shown here.

Observed prompt:

[PRODUCT] reassembling in midair from scattered pieces, reverse-disintegration effect, mechanical precision, each component suspended at a different depth, dark void background, high-concept product advertising, cinematic VFX.

Output
_{Suspended product reassembly concept}

Case 4: Bulk Amazon Listing Generation

Note

Use this when you need to generate listings in bulk across multiple SKUs of the same category. Per-language batch runs are not yet supported.

Case 5: Bulk Product Main Image Replacement

Input: Upload the source product image and the target listing image set.

Prompt:

Replace the product in image 1 across images 2–7. Only the swapped product changes; keep everything else identical.

Note

One-shot recreation of a viral listing — swap the product without redesigning the layout. Useful when many SKUs reuse the same creative style.

Example 1: Globe Ambient Lamp (swapped into the mushroom-lamp viral set)

Source product	Viral listing set (to be replaced)
_{Globe ambient lamp}	_{Original main 1}	_{Original main 2}	_{Original main 3}	_{Original main 4}	_{Original main 5}	_{Original main 6}
	Replaced listing set
	_{Replaced 1}	_{Replaced 2}	_{Replaced 3}	_{Replaced 4}	_{Replaced 5}	_{Replaced 6}

Example 2: Wooden Bedside Table (swapped into the red cabinet viral set)

Source product	Viral listing set (to be replaced)
_{Wooden bedside table}	_{Original main 1}	_{Original main 2}	_{Original main 3}	_{Original main 4}
	Replaced listing set
	_{Replaced 1}	_{Replaced 2}	_{Replaced 3}	_{Replaced 4}

Case 6: Bulk Listing Translation

Input: Upload all listing images that contain text.

Prompt:

Translate the text in all images into {target_language}.

Note

Replace {target_language} before use. Example: "Translate the text in all images into English." Batch-translates listing copy across languages while preserving the original layout — perfect for one-click global rollout.

Example 1: Cabinet Listing (ZH → EN)

Original
_{Original 1}	_{Original 2}	_{Original 3}	_{Original 4}
Translated
_{Translated 1}	_{Translated 2}	_{Translated 3}	_{Translated 4}

Example 2: Cabinet Listing (multilingual batch translation)

Original
_{Original 1}	_{Original 2}	_{Original 3}	_{Original 4}
Translated
_{Translated 1}	_{Translated 2}	_{Translated 3}	_{Translated 4}

👗 Model Try-On and Apparel Prompts

Case 1: Apparel Try-On Image Set

Prompt:

This is a {apparel_type}. Generate a try-on image set for {country}.

Note

Replace {apparel_type} and {country} before use. Example: "This is a dress. Generate a try-on image set for the US." Works for dresses, tops, outerwear, kidswear, and any apparel category that needs commercial try-on display.

Example 1: Green Floral Dress

Input	Output
_{Clean background product}	_{Model front}	_Multi-angle	_{Detail close-up}
_{Clean background product}	_{Full styling}	_{Lifestyle scene}	_{Alternate angle}

Example 2: Indian Sari Traditional Dress

Input	Output
_{Clean background product}	_{Model front}	_{Back / turn}	_Detail
_{Clean background product}	_{Full styling}	_{Lifestyle scene}	_{Alternate angle}

Case 2: Model and Pet Try-On

Input: Upload the apparel or accessory image.

Prompt:

Generate a model wearing this garment.
Generate a pet dog wearing this garment.

Note

This template covers both human-model try-on and pet try-on, suitable for apparel, cosplay props, and pet fashion creatives that need to showcase the same garment on different subjects.

Input	Output	Input	Output	Input	Output
_{#LOVE printed tee}	_{Male model wearing}	_{Cycling tee}	_{Male model wearing}	_{Half-zip knit}	_{Female model wearing}

Case 3: Model Pose Expansion

Input: Upload the original model-with-apparel image.

Prompt:

Expand the model into more poses.

Note

Generates side, 45° turn, walking, and other pose variations to enrich an apparel asset set without re-shooting.

Example 1: #LOVE Printed Tee

Input	Output
_{Original model}	_{Front stance}	_Side	_{45° turn}	_Walking

Example 2: Grey Half-Zip Pullover

Input	Output
_{Original model}	_{Front stance}	_Side	_45°	_Walking

Case 4: Model Swap with Appearance Control

Input: Upload the original model-with-apparel image.

Prompt:

Replace the model with a {ethnicity} model, {hair_description}, while keeping the outfit unchanged.
Replace the figure with a {ethnicity} model, {skin_tone} skin, {hair_color} hair, while keeping the outfit unchanged.

Note

Replace {ethnicity}, {hair_description}, {skin_tone}, and {hair_color} before use. Example: "Replace the model with a Black model, dreadlocks hairstyle, while keeping the outfit unchanged" or "Replace the figure with a Caucasian female model, fair skin, blonde hair, while keeping the outfit unchanged." Helps you localize fashion creatives for different audiences without re-shooting.

Example 1: #LOVE Printed Tee

Input	Output
_{Original model}	_{Swapped model v1}	_{Swapped model v2}

Example 2: Grey Half-Zip Pullover

Input	Output
_{Original model}	_{Swapped model}

Case 5: Reference-Based Model Replacement

Input: Upload the original model image and the reference model image.

Prompt:

Replace the model in image 1 with the model in image 2, keeping the same pose.

Note

Use this when you want to lock a specific model identity across multiple creative assets while reusing the original outfit and pose.

Example 1: Female Model Swap

Original model	Reference model	Output
_{Original model}	_{Reference model}	_{Swap result}

Example 2: Male Model Swap

Original model	Reference model	Output
_{Original model}	_{Reference model}	_{Swap result}

Case 6: Background Concepts for Apparel Sales

Prompt:

Generate suitable background images for {apparel_type} sales display. Provide several indoor and outdoor options to choose from.

Note

Replace {apparel_type} before use. Example: "Generate suitable background images for kidswear sales display. Provide several indoor and outdoor options to choose from." Useful when you separate scene ideation from model generation in your workflow.

Example 1: Beige Knit Pullover

Input	Output
_{Original model}	_{Indoor minimal}	_{Outdoor natural 1}	_{Outdoor natural 2}
_{Original model}	_{City street}	_{At-home scene}	_{Casual scene}

Example 2: Dark Blue Cable Knit

Input	Output
_{Original model}	_{Indoor scene 1}	_{Outdoor scene 1}	_{Outdoor scene 2}
_{Original model}	_{City street}	_{Home environment}	_{Other scene}

Case 7: Accessory Try-On

Input: Upload the accessory product image.

Prompt:

A model wearing this {accessory_type}, {additional_requirements}.

Note

Replace {accessory_type} and {additional_requirements} before use. Example: "A model wearing this necklace, face not shown" or "A model wearing these earrings." Works for necklaces, earrings, hats, scarves, and other wearable accessories.

Input	Output	Input	Output
_{Jade pendant necklace}	_{Worn on model}	_{Gold hoop earrings}	_{Worn on model}

Case 8: Nail Application

Input: Upload the nail design reference image.

Prompt:

Apply these nails to hands.

Note

Useful for press-on nails, gel-polish concepts, salon design previews, and nail product marketing creatives.

Input	Output	Input	Output
_{Pink nail samples}	_{Applied to hand}	_{Multi-color nail samples}	_{Applied to hand}

Case 9: Model Shoe Try-On

Input: Upload the shoe product image.

Prompt:

A model wearing the shoes from the image, {styling_description}.

Note

Replace {styling_description} before use. Example: "A model wearing the shoes from the image, with a brown short skirt." Works for sneakers, heels, boots, and seasonal footwear creatives.

Input	Output	Input	Output
_{Black leather shoes}	_{Male styling}	_{Black high heels}	_{Female styling}

📦 Product Showcase Prompts

Case 1: Clean White Background Product Retouch

Input: Upload a high-quality product photo.

Prompt:

Generate a clean white-background retouched image.

Note

The better the source quality, the more accurate the result. High-quality inputs typically yield more faithful product preservation and cleaner ecommerce-grade output.

Input	Output	Input	Output
_{In-scene product photo}	_{White-background retouch}	_{Raw product photo}	_{White-background retouch}

Case 2: Multi-Angle Product Views

Input: Upload the product reference image.

Prompt:

Generate multi-angle product images.

Note

Useful for product cards, carousels, feature breakdowns, and angle-specific detail pages.

Example 1: Tan Hiking Boots

Input	Output
_{Product photo}	_Front	_{45° side}	_Back
_{Product photo}	_Sole	_Top-down	_{Other angle}

Example 2: Hiking Boots (alternate model)

Input	Output
_{Product photo}	_Front	_Side	_Back
_{Product photo}	_{Sole detail}	_45°	_{Alt angle}

Case 3: Material and Color Variants

Input: Upload the product or apparel image to edit.

Prompt:

Modify the material. Change the figure's clothing into: {material1}, {material2}, and {material3}.

Note

Replace {material1}, {material2}, and {material3} before use. Example: "Modify the material. Change the figure's clothing into: light grey suede, dark brown denim, and burgundy corduroy." Works for fabric replacement, material testing, seasonal recoloring, and concept exploration.

Example 1: Green Jacket Material Variants

Input	Output
_{Original male model}	_{Variant 1}	_{Variant 2}	_{Variant 3}

Example 2: Black Leather Jacket Material Variants

Input	Output
_{Original male model}	_{Variant 1}	_{Variant 2}	_{Variant 3}

Case 4: Smart Background Generation

Input: Upload the product image.

Prompt:

Generate suitable product images.

Note

Use this when you want the model to infer the most fitting commercial product backgrounds without detailed art direction.

Example 1: Snack Box

Input	Output
_{Product photo}	_{Smart background 1}	_{Smart background 2}
_{Product photo}	_{Smart background 3}	_{Smart background 4}

Example 2: Skincare Serum

Input	Output
_{Product photo}	_{Smart background 1}	_{Smart background 2}
_{Product photo}	_{Smart background 3}	_{Smart background 4}

Case 5: Background Replacement with Custom Description

Input: Upload the product image.

Prompt:

Replace the background of the product in the image: {background_description}.

Note

Replace {background_description} before use. Example: "Replace the background of the product in the image: a close-up of the perfume bottle resting on a smooth white stone, surrounded by soft white and pale-blue flowers, with a calm blue lake and clear-sky mountains in the distance — fresh, airy, and slightly cool aesthetic." Useful when you want full art direction over the new scene.

Input	Output	Input	Output
_{Product photo}	_{Custom-described background}	_{Product photo}	_{Custom-described background}

Example 3: Solaris Sunscreen Shield Ad (Community, by @iamrealsnow)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet only shared the output images.

Observed prompt:

SUNSCREEN AD, "THE INVISIBLE SHIELD"

Luxury skincare advertising masterpiece, a colossal premium sunscreen bottle standing on a pristine tropical shoreline at golden hour, powerful beams of sunlight crashing down from the sky and splitting apart upon contact with a transparent protective energy dome radiating from the sunscreen, millions of sparkling UV particles dissolving into golden dust before reaching flawless skin, crystal clear ocean reflections, flowing water suspended in mid air around the product, microscopic droplets catching cinematic sunlight, ultra realistic textures revealing every detail of the bottle surface, luxury beauty campaign aesthetics, dramatic volumetric lighting, glowing atmospheric haze, premium white and gold color palette, futuristic protection technology visualized as elegant light waves, hyper detailed environment, commercial photography perfection, award winning advertising design, photorealistic rendering, 16K ultra resolution, global skincare brand campaign, masterpiece quality.

Text Overlay:
SUNSCREEN

Tagline:
"Protect Every Ray. Reveal Every Glow.

Output
_{Protection dome concept}	_{Golden-hour hero variant}

Example 4: Themed Skeleton Watch Grid (Community, by @Gdgtify)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet only shared the output image.

Observed prompt:

2x2 grid, do this for 4 stunning themes, 16:9 [luxury_watch_cad_system] base_chassis: "hyper-luxury skeletonized tourbillon watch" conceptual_theme: "[$theme, e.g., the samurai shogun]" <mechanism_generation> - ai inference: do not just paint the watch. build the gears, hands, and face using elements from the [$theme]. - example: if ocean, the gears look like coral patterns and the hands are trident spears. if space, the center is a glowing meteorite. - materials: sapphire crystal, forged carbon, rose gold, glowing tritium tubes. </mechanism_generation> <photography> a macro product shot of the watch face. extreme close-up. lighting: dark background, sharp rim lighting catching the metallic edges of the gears. resolution: 8k, razor-sharp focus on the tourbillon movement. vibe: impossibly expensive, intricately engineered, a masterpiece of micro-mechanics. </photography>

Output
_{Four-theme luxury watch grid}

Example 5: Minimalist Serum Still Life (Community, by @iamaiistudio)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet shared the prompt and multiple output variations; the clean hero still life is shown here.

Observed prompt:

Minimalist studio product photography, a small transparent glass facial oil dropper bottle with a black rubber pipette cap, containing pale pink serum with suspended dried pink floral elements, centered on a natural raw wooden block with visible grain and split texture. Tall matte white skincare box on the left labeled "HUILE ÉCLAT VISAGE" with clean black typography and subtle logo near the bottom. Clear cylindrical glass vase on the right filled with water and thin stems of dried pink gypsophila extending upward. Composition rests on a smooth matte pastel pink surface against a matching seamless pink studio background. Strong directional soft light from the left casts long natural-style shadows of the flowers onto the background, with gentle highlights on the glass, subtle reflections on the serum bottle, and soft texture on the wooden block. Straight-on tabletop camera angle, all objects in sharp focus. Color palette: blush pink, soft rose, warm light wood, clean white, transparent glass. Premium Scandinavian minimalist skincare aesthetic, ultra-realistic, studio-grade.

Output
_{Minimalist serum still life}

Case 6: Reference-Based Product Background

Input: Upload the product image and a reference background image.

Prompt:

Image 1 is my {product}. Generate the new scene using the background style from image 2.

Note

Replace {product} before use. Example: "Image 1 is my ring. Generate the new scene using the background style from image 2." Useful for aesthetic matching, brand consistency, and category-specific visual references.

Example 1: Pink Retro Car

Product	Reference background	Output
_{Pink retro car}	_{Green forest reference}	_{Forest tabletop comp}

Example 2: Red Lipstick

Product	Reference background	Output
_{Red lipstick}	_{Black-gold marble reference}	_{Marble tabletop comp}

🤝 Product Interaction Prompts

Case 1: Handheld Product Close-Up

Input: Upload the product image.

Prompt:

A hand is holding this product, {scene_description}, keep the product color unchanged.
{action_description}

Note

Replace {scene_description} and {action_description} before use. Example: "A hand is holding this product, fashion editorial backdrop, keep the product color unchanged" or "Hand holding an electric drill, drilling into wood." Works for beauty, tools, gadgets, packaged goods, and tactile product demos.

Input	Output	Input	Output
_{Pink lip gloss}	_{Hand-held lip gloss close-up}	_{Wooden-handle mop}	_{Hand-held mop cleaning floor}

Case 2: Generated Human Interaction Scene

Input: Upload the product scene or product image.

Prompt:

Add a human model to this product scene. The model should be {person_description}, in a {pose}, performing {action}.

Note

Replace {person_description}, {pose}, and {action} before use. Example: "Add a human model to this product scene. The model should be a young girl, standing pose, holding a doll" or "Add a human model to this product scene, holding a mop." Useful for showcasing scale, usage, emotion, and audience fit.

Input	Output	Input	Output
_{CNY red arch scene}	_{Qipao model entering}	_{Red & silver headphones}	_{Female model wearing headphones}

Example 3: Luxury Product Editorial Poster (Community, by @iamaiistudio)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet only shared the output image.

Observed prompt:

[attach a luxury product photo]
BRAND: [enter your brand name]

Design a single high-end editorial poster. Preserve the product exactly as shown, unaltered. Identify the brand and position its authentic official logo with correct proportions and a subtle print texture in the most fitting location. Include a complementary attractive model naturally wearing or holding the product. Deep rich colors, HDR quality, premium studio lighting, clean minimal background, ample negative space, ultra-sharp 8K output. No additional text or watermarks.

Output
_{Luxury editorial poster output}

Case 3: Reference Character Product Interaction

Input: Upload the product scene image and a reference character image.

Prompt:

Add the human model from image 2 into the product scene from image 1, with this action: {action_description}.

Note

Replace {action_description} before use. Example: "Add the human model from image 2 into the product scene from image 1, with this action: model lying on the bed." Useful when you need consistent character identity across multiple creative assets.

Example 1: Bedroom King-Size Scene + Male Model

Product scene	Reference character	Output
_{Bedroom king-size scene}	_{Male model in white tee}	_{Male model reclining on bed}

Example 2: Home Rocking-Chair Scene + Male Model

Product scene	Reference character	Output
_{Home rocking-chair scene}	_{Male model in white shirt}	_{Male model in rocking chair}

Example 3: Reference Shopper Product Scene (Community, by @mehvishs25)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet shared the prompt and output images, showing how a preserved reference identity can be moved into a retail product-comparison scene.

Observed prompt:

Photorealistic candid lifestyle photograph of the same woman as the reference image, preserving her exact facial identity, facial proportions, skin tone, and recognizable features. She is casually shopping inside a modern convenience store, captured in an authentic everyday moment rather than a posed photoshoot.

Medium-full body side-angle composition. She stands in a brightly illuminated aisle, comparing two beauty or wellness products while reading packaging details with a focused expression. One hand holds a sleek bottle close to her face while the other carries a second product she is considering.

Her dark hair is styled in a relaxed high bun with a few loose strands framing her face. Slim gold hoop earrings, subtle glossy lips, natural makeup, and oversized black sunglasses resting on her head.

Outfit: a bold, fitted black off-shoulder ribbed top paired with a charcoal gray mini skort and sheer black tights, creating a chic city-girl aesthetic. She carries a large luxury-inspired monogram shoulder bag and wears black ankle boots.

Modern convenience store environment in Southeast Asia with narrow aisles, brightly stocked shelves filled with colorful skincare, beverages, snacks, and household essentials. Fluorescent ceiling panels cast soft, diffused light across the scene.

Natural candid body language, realistic posture, genuine shopping behavior, authentic retail atmosphere, unposed documentary aesthetic. Soft depth of field with subtle background blur while keeping the subject sharply detailed.

Shot with a 35mm lens at eye level, ultra-realistic RAW photography quality, realistic skin texture, accurate fabric details, natural colors, soft shadows, crisp focus, 4K detail.

Vertical 4:5 composition, subject positioned slightly left of center, facing the shelves on the right side of the frame, immersive everyday shopping scene, contemporary urban lifestyle editorial feel without looking staged.

Output
_{Reference shopper comparison scene}

Case 4: Pet and Product Interaction

Input: Upload the product image.

Prompt:

Generate a {pet} interacting with the {product} from the image.

Note

Replace {pet} and {product} before use. Example: "Generate a cat interacting with the toy from the image" or "Generate a puppy interacting with the toy from the image." Especially useful for toy, pet accessory, home, and playful lifestyle product creatives.

Input	Output	Input	Output
_{Orange dog chew ball}	_{Puppy biting ball}	_{Hamster wheel}	_{Hamster in wheel}

🛍️ Social Commerce Prompts

Case 1: Fashion Studio Sales Set

Prompt:

Generate a model sales image set, background is a fashion studio.

Note

Useful for showroom-style apparel marketing, boutique visual merchandising, and studio-shoot social content.

Example 1: Blue Hoodie

Input	Output
_{Clean background product}	_{Studio shoot 1}	_{Studio shoot 2}
_{Clean background product}	_{Studio shoot 3}	_{Studio shoot 4}

Example 2: Pink Floral Dress

Input	Output
_{Clean background product}	_{Studio shoot 1}	_{Studio shoot 2}
_{Clean background product}	_{Studio shoot 3}	_{Studio shoot 4}

Example 3: Football Jersey Editorial Set (Community, by @Taaruk_)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet only shared the output images.

Observed prompt:

Luxury sports fashion campaign, Cristiano Ronaldo inspired editorial photoshoot, standing confidently in a minimalist beige studio, wearing a black Portugal national team jersey with a bold red textured vertical stripe, athletic physique, hands behind back, serious expression, premium football lifestyle aesthetic, high-fashion sportswear photography, clean neutral backdrop, soft cinematic lighting, luxury magazine cover style, Nike-inspired campaign, ultra realistic skin texture, sharp focus, professional studio photography, muted earth-tone color palette, modern luxury branding, 85mm lens, shallow depth of field, photorealistic, 8K, editorial masterpiece.

Additional shots:
1. Full-body standing portrait, centered composition, confident pose.
2. Close-up chest detail shot highlighting jersey texture, number 7, luxury silver chain, premium fabric details.
3. Seated portrait on modern black designer chair, relaxed yet powerful posture, fashion campaign aesthetic.
4. Product display shot featuring jersey draped over chair with matching tracksuit and sneakers, luxury sports catalog photography.

Style keywords:
minimalist luxury, football icon, premium sportswear campaign, modern editorial, Vogue Men, GQ Sports, cinematic studio lighting, high-end fashion photography, realistic textures, luxury athlete branding, clean composition, commercial advertising quality.

Output
_{Studio portrait}	_{Jersey detail close-up}
_{Seated campaign portrait}	_{Product display layout}

Case 2: Mirror Outfit Sales Video

Prompt:

Generate a mirror outfit sales video.

Note

This template is ideal for creator-style fashion content, UGC aesthetics, and casual social-commerce sales formats.

Case 3: Gift Box Styling for Apparel

Input: Upload the apparel image.

Prompt:

Place the apparel into a gift box.

Note

Useful for holiday gift marketing, premium packaging mockups, and seasonal-themed product creatives.

Input	Output	Input	Output
_{Color-splatter tee}	_{Gift box result}	_{Red kids' down jacket}	_{Gift box result}

Case 4: Fresh Food Sales Scene Set

Prompt:

Generate a fresh-food scene set for {food_name}.

Note

Replace {food_name} before use. Example: "Generate a fresh-food scene set for pomelo." Useful for fruits, agricultural produce, boxed foods, specialty ingredients, and premium fresh-food displays.

Example 1: Watermelon

Input	Output
_{Product photo}	_{Fresh-food scene 1}	_{Fresh-food scene 2}
_{Product photo}	_{Fresh-food scene 3}	_{Fresh-food scene 4}

Example 2: Grapes

Input	Output
_{Product photo}	_{Fresh-food scene 1}	_{Fresh-food scene 2}	_{Fresh-food scene 3}
_{Product photo}	_{Fresh-food scene 4}	_{Fresh-food scene 5}

Case 5: General Food Sales Creative Set

Prompt:

Generate a food sales image set.

Note

A flexible base template for snack brands, packaged foods, fresh-food products, and general social-commerce food content.

Example 1: Popcorn

Input	Output
_{Product photo}	_{Sales image 1}	_{Sales image 2}	_{Sales image 3}
_{Product photo}	_{Sales image 4}	_{Sales image 5}

Example 2: Candied Fruit Skewers

Input	Output
_{Product photo}	_{Sales image 1}	_{Sales image 2}
_{Product photo}	_{Sales image 3}	_{Sales image 4}

Example 3: Obsidian Coffee Brand Campaign (Community, by @iamaiistudio)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet shared the prompt and output image.

Observed prompt:

Generate four cohesive high-end realistic editorial visuals for OBSIDIAN coffee brand. Cinematic, dark, mature aesthetic inspired by luxury sportswear and premium coffee advertising. Studio lighting that's dramatic yet controlled, photorealistic textures, clean compositional layout. Shot 1: Hero brand poster featuring 'OBSIDIAN' lettering with an artful coffee display - steam rising, beans scattered. Shot 2: Full product range - coffee bags, cans, and capsules arranged together. Shot 3: Tight packaging detail with tagline 'Coffee for grown-ups who chase flavor.' Shot 4: Lifestyle close-up of a steaming cup. Ultra-polished finish, crisp realistic materials, unified brand identity, no fantastical or surreal elements

Output
_{Four-shot coffee campaign board}

Example 4: Oatmilk Brownie Bottle Poster (Community, by @Shorelyn_)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet shared multiple images, and the matching hero poster is shown here.

Observed prompt:

Premium commercial 3D advertising poster, vertical 4:5 format.

Hero product perfectly centered: Starbucks Frappuccino Oatmilk Dark Chocolate Brownie bottle, floating above a glossy sculptural pedestal.

Massive luxury typography in the background:

"INDULGE THE DARK SIDE"

oversized bold editorial font, partially hidden behind the bottle, layered depth effect.

Additional typography:

"Rich Dark Chocolate. Smooth Oatmilk."

Small premium copy:

"Every Sip Feels Like Dessert."

Environment inspired by the product itself:

flowing ribbons of creamy oatmilk, glossy dark chocolate streams, floating brownie cubes, oat flakes, purple botanical leaves from the packaging design.

Color palette extracted from the label:

deep royal purple, cream beige, dark chocolate brown, warm oat tones.

Luxury graphic design elements:

editorial typography, floating sticker labels, minimal barcode, micro typography, premium packaging callouts, subtle geometric shapes, modern commercial layout.

Studio cinematic lighting, soft key light, purple rim light, volumetric glow, floating particles, realistic glass reflections, hyper-realistic PBR materials, ultra-sharp label readability, shallow depth of field, advertising campaign quality, Behance featured project quality, 8K commercial render, high-end beverage branding, premium food photography aesthetic.

Output
_{Chocolate oatmilk hero poster}

Example 5: Velvet Bite Chocolate Campaign (Community, by @iamaiistudio)

Tip

Community reference from a recent public GPT Image 2 post. The source tweet shared the prompt and multiple campaign frames; one hero square ad is shown here.

Observed prompt:

Design three premium square product ads for a fictional luxury chocolate brand named Velvet Bite.

Each ad features a dark chocolate bar with consistent brand packaging: matte black wrapper, gold foil accents, elegant serif typography, and high-end editorial-style photography.

Give each ad its own distinct flavor world through a unique color palette, props, lighting, and atmosphere:

Blood Orange Noir
Dark chocolate with citrus notes, warm orange tones, dramatic shadows, gold foil detail, and an opulent luxury aesthetic.

Salted Pistachio Muse
Dark chocolate with pistachio, soft cream-toned background, green accent elements, graceful and refined composition.

Raspberry Ember
Dark chocolate with raspberry, deep crimson backdrop, moody atmospheric lighting, rich and indulgent feel.

Every ad must include: a strong bold headline, brief supporting copy, the Velvet Bite bar as the hero product, flavor-matched ingredients styled around the bar, and a CTA reading "Shop the drop."

The overall campaign should feel polished, cohesive, and premium, with all three ads clearly belonging to the same brand identity.

Output
_{Luxury chocolate campaign poster}

🚀 Batch Generation via Evolink API

For production runs across hundreds of SKUs, multiple languages, or multiple marketplaces, call the Evolink GPT Image 2 API directly instead of running prompts one by one.

Why batch via API:

Bulk listing image generation for an entire catalog
Multilingual translation of all listing assets in a single job
Programmatic model swap, background replacement, multi-angle output
Webhook callbacks integrate with your PIM, ERP, or marketing automation

Prerequisites: Get an API key.

Step 1: Submit a generation task

GPT Image 2 generation is asynchronous — the call returns a task_id that you poll for results.

curl -X POST https://api.evolink.ai/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "Generate an Amazon listing image set for the US in English. The product is a solid wood dining chair, and the selling points are sturdy frame, ergonomic backrest, easy assembly.",
    "image_urls": ["https://your-cdn.example.com/products/chair.jpg"],
    "size": "1:1",
    "resolution": "2K",
    "quality": "high",
    "n": 4
  }'

Field	Required	Notes
`model`	yes	Must be `"gpt-image-2"`
`prompt`	yes	Up to 32,000 characters — paste the prompt from any case in this repo
`image_urls`	optional	1–16 input/reference images (jpeg/png/webp, ≤50 MB each) for image-to-image and editing tasks
`size`	optional	Aspect ratio (`1:1`, `2:3`, `16:9`, …) or pixel dimensions; defaults to `auto`
`resolution`	optional	`1K`, `2K`, or `4K` when `size` is a ratio
`quality`	optional	`low`, `medium`, `high` — affects cost; defaults to `medium`
`n`	optional	1–10 images per call
`callback_url`	optional	HTTPS webhook to receive completion instead of polling

Step 2: Poll the task

curl https://api.evolink.ai/v1/tasks/{task_id} \
  -H "Authorization: Bearer YOUR_API_KEY"

When status becomes "completed", the results array contains the generated image URLs. Generated images are valid for 24 hours — download them promptly.

Optional: webhook callback

Pass "callback_url": "https://your-server.example.com/webhook" in the submission, and Evolink will POST the task result to your endpoint when the task completes — no polling required.

Full API reference: docs.evolink.ai · GPT Image 2 image generation

🙏 Acknowledge

This repository is inspired by excellent open-source prompt collections and community-shared ecommerce workflows.

Thanks to all creators and contributors who openly share their work and make these case studies possible.

@EvoLinkAI

We cannot guarantee that every case is attributed to its original creator. Please contact us if any correction is needed and we will update promptly.

If you have more interesting prompt cases to share, feel free to reach out and help us extend the Evolink prompt library.

Name		Name	Last commit message	Last commit date
Latest commit History 17 Commits
.github		.github
data		data
e-commerce		e-commerce
images		images
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
README_de.md		README_de.md
README_es.md		README_es.md
README_fr.md		README_fr.md
README_ja.md		README_ja.md
README_ko.md		README_ko.md
README_pt.md		README_pt.md
README_ru.md		README_ru.md
README_tr.md		README_tr.md
README_zh-CN.md		README_zh-CN.md
README_zh-TW.md		README_zh-TW.md
review.md		review.md

Folders and files

Latest commit

History

Repository files navigation

EvoLink Quick Start

How to Use GPT Image 2 for E-commerce

🍌 Introduction

How to Use These Prompts

📑 Menu

🖼️ Ecommerce Listing Image Prompts

Case 1: Marketplace Listing Image Set

Example 1: Solid Wood Dining Chair

Example 2: Sport Water Bottle

Example 3: Sport T-shirt

Case 2: Multi-Country Listing Image Set

Example 1: Purple Polka-Dot Dress

Example 2: Cordless Vacuum

Example 3: Christmas-Theme Phone Case

Case 3: A+ Product Detail Visuals

Example 1: Premium Tea Gift Box

Example 2: Natural Glue Stick

Example 3: Bubble Tea Drink

Example 4: Sony A7 Exploded View (Community, by @iamaiistudio)

Example 5: Architectural Product Catalog Page (Community, by @iamaiistudio)

Example 6: Reverse Disintegration Product Ad (Community, by @iamaiistudio)

Case 4: Bulk Amazon Listing Generation

Case 5: Bulk Product Main Image Replacement

Example 1: Globe Ambient Lamp (swapped into the mushroom-lamp viral set)

Example 2: Wooden Bedside Table (swapped into the red cabinet viral set)

Case 6: Bulk Listing Translation

Example 1: Cabinet Listing (ZH → EN)

Example 2: Cabinet Listing (multilingual batch translation)

👗 Model Try-On and Apparel Prompts

Case 1: Apparel Try-On Image Set

Example 1: Green Floral Dress

Example 2: Indian Sari Traditional Dress

Case 2: Model and Pet Try-On

Case 3: Model Pose Expansion

Example 1: #LOVE Printed Tee

Example 2: Grey Half-Zip Pullover

Case 4: Model Swap with Appearance Control

Example 1: #LOVE Printed Tee

Example 2: Grey Half-Zip Pullover

Case 5: Reference-Based Model Replacement

Example 1: Female Model Swap

Example 2: Male Model Swap

Case 6: Background Concepts for Apparel Sales

Example 1: Beige Knit Pullover

Example 2: Dark Blue Cable Knit

Case 7: Accessory Try-On

Case 8: Nail Application

Case 9: Model Shoe Try-On

📦 Product Showcase Prompts

Case 1: Clean White Background Product Retouch

Case 2: Multi-Angle Product Views

Example 1: Tan Hiking Boots

Example 2: Hiking Boots (alternate model)

Case 3: Material and Color Variants

Example 1: Green Jacket Material Variants

Example 2: Black Leather Jacket Material Variants

Case 4: Smart Background Generation

Example 1: Snack Box

Example 2: Skincare Serum

Case 5: Background Replacement with Custom Description

Example 3: Solaris Sunscreen Shield Ad (Community, by @iamrealsnow)

Example 4: Themed Skeleton Watch Grid (Community, by @Gdgtify)

Example 5: Minimalist Serum Still Life (Community, by @iamaiistudio)

Case 6: Reference-Based Product Background

Example 1: Pink Retro Car

Example 2: Red Lipstick

🤝 Product Interaction Prompts

Case 1: Handheld Product Close-Up

Case 2: Generated Human Interaction Scene

Example 3: Luxury Product Editorial Poster (Community, by @iamaiistudio)

Case 3: Reference Character Product Interaction

Example 1: Bedroom King-Size Scene + Male Model

Example 2: Home Rocking-Chair Scene + Male Model

Example 3: Reference Shopper Product Scene (Community, by @mehvishs25)

Case 4: Pet and Product Interaction

🛍️ Social Commerce Prompts

Packages