Nano Banana 2 Flat Lay Product Photo Prompts for Controlled Restyling

Quick answer

Use the original flat lay as the product source of truth, lock each object that must remain identical, then change one styling variable at a time: spacing, surface, shadow softness, or prop arrangement. Keep price, badges, dimensions, claims, and annotations out of the generated image and add them in an editable product template.

Who this guide is for

Ecommerce sellers, catalog designers, and social teams who already have usable top-down product photography but need cleaner spacing, seasonal surfaces, or channel-specific crops.

Recommended model

Use case	Recommended model	Why
This workflow	Nano Banana 2 in image-to-image mode	The reference-led workflow is suited to rearranging a photographed arrangement while preserving the exact SKU, labels, materials, and relative product scale.

AIBase is an independent creative platform. Model names are shown only to identify supported underlying technologies and workflow choices.

Prompt template

Use the uploaded authorized flat-lay photo as the exact product source. Preserve every product silhouette, SKU, colorway, label position, cap or closure, material texture, quantity, logo zone, and relative scale. Restyle only the arrangement into a [grid / diagonal / centered hero] on [surface]. Use [one light direction], soft contact shadows, and [negative-space position] for an editable selling zone. Keep the camera directly overhead and the frame aligned. Do not add products, hands, text, price, badge, claim, measurement, watermark, or changed package contents.

Step-by-step workflow

1. Start from the sharpest top-down source and mark each real product, included accessory, and label that must remain unchanged.
2. Decide the destination channel and safe margins before restyling so the hero product and selling zone survive the crop.
3. Write the protected-object list first, then specify only one styling change: spacing, surface, props, or light.
4. Generate the full rearranged scene once before requesting color, seasonal, or crop variants.
5. Compare product proportions, label legibility, edge cleanliness, reflections, and shadow direction against the source.
6. Composite approved text, dimensions, ingredients, price, promotions, and disclosures in an editable catalog template.
7. Retest mobile thumbnails, zoom detail, accessibility, and channel crops before publishing the final asset.

Example prompt variants

• Keep both approved jars, lids, labels, spatula, and quantities unchanged; restyle into a three-column grid on pale linen, top-down camera, soft shadows, and a blank lower-third selling zone.
• Keep the exact bottle, dropper, cap, and label geometry; move it to the right third of a warm stone surface, preserve overhead perspective, and reserve the left half for editable copy.
• Keep all four approved colorways and package proportions; create a tidy diagonal arrangement on a cool gray sweep with equal spacing and no added props or text.

Quality checklist

• Every SKU, colorway, package quantity, closure, label, and logo zone matches the authorized source.
• Product proportions and contact shadows remain physically credible in the new arrangement.
• Label edges, transparent materials, threads, and small controls stay legible at zoom.
• The selling zone survives desktop, mobile, thumbnail, and marketplace crops.
• Claims, dimensions, price, inventory, badges, and accessibility text remain editable outside the image.

Mistakes to avoid

• Letting the model quietly add accessories, substitute packaging, or change the visible quantity.
• Requesting several styling changes in one pass, which makes the best output difficult to reproduce.
• Hardcoding prices, claims, promotions, or measurements into generated pixels.
• Publishing a restyled arrangement that no longer represents the actual product or included items.

Related AIBase pages

• Restyle a flat lay with Nano Banana 2
• Use reference-preserving prompt patterns
• Compare flat lay model workflows

Practical next step

Open the most relevant AIBase generator, run one narrow prompt first, and save the best result before adding more constraints. The fastest way to improve output quality is to compare one variable at a time: subject, camera, background, lighting, then final polish.