GPT Image 2 Product Photo Prompts for Packshots and Campaigns

A useful GPT Image 2 product photo prompt is a production brief. It defines the product, camera, surface, light, background, layout, and non-negotiable details. Style adjectives come last. This structure makes outputs easier to judge and makes revisions less destructive.

The examples below are prompt-writing patterns for GPT Image 2 workflows. Model availability in Genzy depends on the image providers enabled for the workspace.

Start with the product identity block

Write down what cannot change: object category, silhouette, dimensions, material, finish, color, closure, label location, and distinctive construction. For a cosmetic bottle, that might be: “Tall rectangular 50 ml glass fragrance bottle, thick transparent base, matte black cylindrical cap, centered cream label with generous margins.”

Avoid asking the model to infer critical geometry from brand mood. “Premium perfume” can produce many incompatible shapes; the identity block defines the one you want.

Add a packshot composition

For a clean ecommerce image, use a prompt such as: “Create a front-facing studio packshot of the described bottle, centered on a seamless warm-white background. Camera at product height with a 70 mm commercial product-photo look. Soft large key light from upper left, gentle fill from the right, subtle grounded shadow, accurate transparent glass, crisp edges. Preserve the exact bottle proportions and cap alignment. No props, hands, or decorative text.”

The last sentence removes common distractions without turning the prompt into a long negative list.

Write for label and text safety

If exact packaging text matters, provide a clean reference and state what must be preserved. Keep the requested copy short and specify placement: “Keep the existing label artwork unchanged and centered; do not invent additional words or badges.” After generation, inspect spelling, letter shape, line breaks, and logo proportions at full size. Treat packaging text as a required review item, not something guaranteed by an attractive thumbnail.

For a concept image without final artwork, request a blank label area and add production typography later in a design tool.

Create a campaign scene without losing the product

Use the packshot as an approved anchor, then add the environment around it: “Preserve the exact bottle, cap, and label from the reference. Place it on a dark green stone pedestal in a humid botanical greenhouse at dawn. Condensation appears on the pedestal, not over the label. A narrow beam of light creates a controlled highlight on the bottle edge. Leave clean negative space in the upper right for campaign copy.”

Notice that product preservation comes before the new scene. The prompt also protects the label and reserves layout space for a real marketing use.

Control reflective and transparent materials

Name how the material behaves under light. Glass needs transmission, refraction, clean edge highlights, and a believable contact shadow. Brushed metal needs directional texture and controlled reflections. Glossy plastic needs a broad softbox reflection rather than random bright spots. Fabric needs weave, edge softness, and appropriate folds.

Instead of “realistic materials,” write “brushed aluminum with fine vertical grain, one broad rectangular softbox reflection, no mirror-like room reflections.” Specific light behavior is more useful than another realism adjective.

Build a consistent variant set

Approve one composition before asking for multiple colors or flavors. Then change one product attribute at a time while preserving camera height, crop, shadow, background, and label grid. A variant prompt might say: “Keep the approved camera, lighting, bottle geometry, cap, label placement, and background unchanged. Change only the liquid and accent color from amber to pale rose.”

Generate each variant from the same anchor when possible. Compare them in a contact sheet to catch small shifts in height, width, cap size, and label position.

Product photo prompt templates

For an overhead flat lay: “Top-down product arrangement on a matte neutral surface, exact object geometry preserved, balanced spacing, soft directional daylight, restrained shadows, clear empty area on the left for a headline.”

For a dramatic launch image: “Low three-quarter hero angle, product large in frame, controlled dark gradient background, narrow rim light defining the silhouette, one material-appropriate reflection, grounded shadow, no floating objects, final packaging details preserved.”

For a social ad: “Vertical 4:5 crop, product in the lower center, one supporting ingredient element, simple color-block background, strong separation at thumbnail size, upper third left open for copy, no generated text.”

Review against a fixed checklist

Check silhouette, proportions, material, label, readable text, color, cap or closure, shadows, reflections, crop, and negative space. If only one part fails, rewrite that requirement rather than regenerating with a completely new style prompt.

The most efficient GPT Image 2 product workflow moves from identity block to clean packshot, then to campaign environment, then to controlled variants. That order gives every creative image a stable product reference and prevents visual polish from hiding packaging drift.