EmbeddingOptions

Defined in: packages/ai/src/types.ts:3358

Options for embedding generation, as received by adapters. The embed() entry point normalizes a single input item to an array before calling the adapter, so input is always an array here.

Type Parameters#

TProviderOptions#

TProviderOptions extends object = object

Properties#

abortSignal?#

optional abortSignal?: AbortSignal;

Defined in: packages/ai/src/types.ts:3381

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.

dimensions?#

optional dimensions?: number;

Defined in: packages/ai/src/types.ts:3367

Requested output dimensionality. Adapters for models with fixed dimensions throw a clear runtime error when this is set.

input#

input: EmbeddingInputItem[];

Defined in: packages/ai/src/types.ts:3362

The items to embed — one vector per item

logger#

logger: InternalLogger;

Defined in: packages/ai/src/types.ts:3375

Internal logger threaded from the embed() 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:3360

The model to use for embedding generation

modelOptions?#

optional modelOptions?: TProviderOptions;

Defined in: packages/ai/src/types.ts:3369

Model-specific options for embedding generation