Interface: ImageGenerationOptions<TProviderOptions, TSize>

Defined in: packages/ai/src/types.ts:1936

Options for image generation. These are the common options supported across providers.

Type Parameters

TProviderOptions

TProviderOptions extends object = object

TSize

TSize extends string | undefined = string

Properties

abortSignal?

optional abortSignal?: AbortSignal;

Defined in: packages/ai/src/types.ts:1968

Effective abort signal composed by the activity from caller abortSignal and/or timeout. Adapters should forward this to the provider SDK when supported. Request-specific — never store on a global client config.

logger

logger: InternalLogger;

Defined in: packages/ai/src/types.ts:1962

Internal logger threaded from the generateImage() entry point. Adapters must call logger.request() before the SDK call and logger.errors() in catch blocks.

model

model: string;

Defined in: packages/ai/src/types.ts:1941

The model to use for image generation

modelOptions?

optional modelOptions?: TProviderOptions;

Defined in: packages/ai/src/types.ts:1957

Model-specific options for image generation

numberOfImages?

optional numberOfImages?: number;

Defined in: packages/ai/src/types.ts:1953

Number of images to generate (default: 1)

prompt

prompt: MediaPrompt;

Defined in: packages/ai/src/types.ts:1951

Description of the desired image(s): a plain string, or an ordered array of content parts for image-conditioned generation (image-to-image, reference-guided, edit, multi-reference). Media parts may carry metadata.role to disambiguate intent (mask, control, reference, …). Adapters map parts onto the provider-native request — e.g. Gemini multimodal contents, OpenAI images.edit(), fal image_url / mask_url — and throw a clear runtime error for unsupported modalities.