Skip to main content
POST https://api.voidai.app/v1/images/generations Send JSON with authentication. Image routes are unavailable to free-plan accounts. Choose a model advertising this endpoint, such as gpt-image-2; models that only advertise chat/Responses use those routes instead.

​
Request fields

The schema accepts streaming options, but current GPT image catalog flags do not advertise streaming. Treat streaming as model/provider-dependent and use non-streaming unless support is confirmed.

​
Sizes and model constraints

Named sizes accepted by the gateway are 256x256, 512x512, 1024x1024, 1536x1024, 1024x1536, 2048x2048, 2048x1152, 3840x2160, 2160x3840, 4096x4096, and auto. Custom dimensions must be multiples of 16, have no edge above 3840, have an aspect ratio at most 3:1, and contain between 655,360 and 8,294,400 pixels. These custom-size checks do not apply to the named sizes above. Selected providers can impose additional limits. All gpt-image* models reject style and response_format. Use output_format to choose the encoding; read b64_json or url from the result rather than assuming a URL.

​
Response and billing

Non-streaming responses contain created and a data array with url and/or b64_json, and optionally revised_prompt. Images are billed per output image; see billing and current prices.