Skip to main content
← GPT Model name: gpt-image-1.5/edit
Generation type: image_to_image
Endpoint: POST /api/asset_generator/

Pricing

Prices are per output. When a request asks for several outputs, the charge is this price times the number of outputs.

Quick start

The example uses sample media from our CDN. Swap in publicly accessible URLs for your own files.

Input parameters

Send these fields in input_params.
string
required
The prompt for image generationMinimum characters: 2. Maximum characters: 32000.
string[]
required
The URLs of the images to use as a reference for the generation.Minimum items: 1. Maximum items: 10.
string
default:"high"
Quality for the generated imageAllowed values: "low", "medium", "high".
string
default:"auto"
Background for the generated imageAllowed values: "auto", "transparent", "opaque".
string
default:"auto"
Aspect ratio for the generated imageAllowed values: "auto", "1024x1024", "1536x1024", "1024x1536".
integer
default:1
Number of images to generateMinimum: 1. Maximum: 4.
string
default:"png"
Output format for the imagesAllowed values: "jpeg", "png".
string
default:"high"
Input fidelity for the generated imageAllowed values: "low", "high".

Response

The request returns the task with status: "initializing". Poll GET /api/asset_generator/{id}/ until status is done or failed, or pass a webhook_url to be notified. A finished task lists its files in assets and the credits it charged in credits_used.