Product photos - three backgrounds, one item

A product photo series needs a consistent brief. Change the surroundings without changing the item’s material, color or proportions. Text-only generation can develop a concept. For a real listing, use your own product reference and verify that the output accurately represents the item.

Prepare the shot

• Record the features that must not change: color, parts, texture, logo and shape.
• Make a simple catalog background with a soft shadow and clear edges.
• Create a lifestyle setting with one scale cue, without implying extra accessories are included.
• Prepare a detail view and reject invented quality labels.
• Compare the three images together for changes in handles, closures and geometry.
• Export consistent aspect ratios and write alt text describing the product and view.

Review before publication

• Color matches the source.
• Accessories do not misrepresent the package.
• Each image serves a different listing purpose.

API input example

The JSON below is the request body for POST /v1/generate. Use your own accessible image URL where a reference is required.

{
  "model": "gpt-image-2.5",
  "input": {
    "prompt": "A plain ceramic mug on a neutral light gray background, soft shadow, catalog photography, no logo, no text"
  }
}

Model and cost

Use GPT Image 2.5. Its current unit price is $0.0432 per image. Check the pricing page and calculator for your settings. Credits never expire, there is no subscription and failed generations are free. A successful output is billable even if you choose not to publish it.

FAQ

Can text-only generation represent a specific product?

Only after a fidelity review. Text alone does not document the sold item and can lead to invented details.

Can I check the request before a paid run?

Yes. Check the model schema and the cost shown in Studio. Local demo API tests validate request handling but do not prove the visual or audio quality of the real model.

Where are the supported parameters?

Model details and the generation API guide list the supported inputs. Preview mode does not provide a real generated result.