Use it for
Best first use
Use this workflow when one product photo is already approved and the job is to replace the scene for a gallery image, ad base, or catalog refresh.
Product Background Workflow
Replace a product photo background without rewriting the product. Start with one SKU, choose a studio, lifestyle, or seasonal background, then inspect the product before scaling variants.
Short Answer
The best first test is one approved product photo, one background job, and one inspection pass. Keep strict white-background marketplace slots separate from lifestyle, seasonal, and campaign scenes.
Use it for
Use this workflow when one product photo is already approved and the job is to replace the scene for a gallery image, ad base, or catalog refresh.
Avoid
Do not treat lifestyle, seasonal, or textured backgrounds as strict marketplace main-image candidates. Keep those slots separate.
Handoff
The Studio handoff opens Image to Image with GPT Image-2, a square 2K setup, a product source image, and a preservation prompt.
Review
Check label area, outline, material, contact shadow, and unsupported props or claims before generating more background variants.
Workflow
Upload one product image, pick one publishing slot or style, inspect product fidelity, then replace the sample source with your SKU before scaling variants.
Start with one product photo that already shows the item clearly. The source image is the control layer for shape, label, material, and scale.
Ask for a studio surface, lifestyle setting, seasonal colorway, or ad-ready scene. Do not combine every campaign idea in the first run.
Inspect edges, logo area, label readability, contact shadow, and props before generating variants or changing the creative direction.
Background Types
A background that works for a Shopify collection tile may be wrong for an Amazon main image or a text-heavy ad. Decide the slot before the prompt.
First pass
Use for catalog consistency, PDP galleries, collection grids, and retakes where the product should feel clean but not flat.
Context
Use when context helps the shopper understand material, use case, scale, or setting. Keep the product larger than the scene.
Campaign
Use campaign colors, props, or light cues after the product survives a simple studio pass. Avoid holiday clutter over the label.
Wrong slot
Reject lifestyle, seasonal, prop-heavy, or textured scenes for strict main-image slots. Use a white-background review path with no text, watermarks, confusing props, or graphic overlays.
Visual Cases
The useful result is not just a prettier scene. It is a background change that keeps the product inspectable and fits the slot where it will be published.
Control
The label, package edges, and existing shadows give the edit a product-fidelity target.
Catalog
Use this when the catalog needs a cleaner surface without changing packaging or inventing selling claims.
Secondary
The scene adds context around the product. It should stay outside the label area and avoid confusing extra items.
Wrong slot
The product is still clear, but the campaign colors make this wrong for a strict main-image slot. Use it for ads, emails, or secondary gallery context.
First Test
A useful first result changes the scene while keeping the product inspectable. Use these checks before spending credits on more backgrounds.
The silhouette, cap, pouch edge, box corner, bottle pump, or handle should still match the source image.
Readable labels may soften, but the model should not invent new brand names, claims, ingredients, or certifications.
The new background needs a believable grounding shadow. Floating products usually mean the scene changed faster than the product mask.
Reject outputs that add accessories, awards, discount badges, medical claims, or bundled items you did not provide.
Prompts
Keep preservation language in every prompt. Background replacement fails fastest when the prompt asks for a full ad scene before the product mask is stable.
Replace the background with a warm neutral ecommerce studio surface. Preserve the exact product shape, label text, color, material, and proportions. Add a soft contact shadow and subtle light falloff. No new text, no badges, no extra products, no hand interaction, no logo changes.
Place this product in a clean lifestyle background that suggests a modern kitchen counter. Keep the product large, sharp, and unchanged. Add context behind and around the product only. Do not cover the label, add props in front, invent claims, or change packaging.
Create a seasonal campaign background using soft winter light and restrained color accents. Preserve product geometry, label area, material, and scale. Keep decorative elements in the background, not touching the product. No sale badges, no new copy, no unsupported claims.
Studio Handoff
The primary CTA opens GPT Image-2 in image-to-image mode with the product source and preservation prompt already filled in. Replace the sample source with your product photo before generating.
Use when the product visual needs Amazon-specific slot guidance, secondary infographics, or main-image risk checks.
Animate the approved product still into a PDP loop or paid social product clip.
Add campaign copy and readable callouts after the product background is stable.
These answers focus on the practical product-photo job: replacing the scene while keeping the product useful for ecommerce and ads.
Use a clean product photo with the full product visible, clear edges, stable lighting, and enough resolution to read the label. Background replacement is weaker when the product is cropped, blurry, reflective, or covered by hands.
Use a strict white-background workflow for Amazon main-image candidates. This page is better for ecommerce galleries, secondary images, product ads, Shopify-style catalog refreshes, and lifestyle backgrounds where context is allowed.
The preset opens GPT Image-2 in image-to-image mode with a 1:1, 2K setup and a prompt that protects product shape, label area, edges, material, and contact shadow.
Check whether the product outline, logo area, readable label text, proportions, and shadows still match the source. Reject the output if it adds unsupported claims, extra products, badges, or confusing props.