Class: TemporalEmbeddingModel

aisdk.TemporalEmbeddingModel

An embedding model implementation that delegates embedding generation to Temporal activities. This allows workflows to generate embeddings through the Temporal execution model.

The AI SDK integration is an experimental feature; APIs may change without notice.

Implements

• EmbeddingModelV3

Constructors

constructor

• new TemporalEmbeddingModel(modelId, options?): TemporalEmbeddingModel

Parameters

Name	Type
modelId	string
options?	ActivityOptions

Returns

TemporalEmbeddingModel

Properties

maxEmbeddingsPerCall

• Readonly maxEmbeddingsPerCall: undefined

Undefined to let the underlying provider handle chunking, as it knows its own limits.

Implementation of

EmbeddingModelV3.maxEmbeddingsPerCall

---

modelId

• Readonly modelId: string

Implementation of

EmbeddingModelV3.modelId

---

options

• Optional Readonly options: ActivityOptions

---

provider

• Readonly provider: "temporal"

Implementation of

EmbeddingModelV3.provider

---

specificationVersion

• Readonly specificationVersion: "v3"

Implementation of

EmbeddingModelV3.specificationVersion

---

supportsParallelCalls

• Readonly supportsParallelCalls: true

Indicates the underlying embedding model API can handle concurrent requests. Set to true since we delegate to the actual provider which manages its own concurrency.

Implementation of

EmbeddingModelV3.supportsParallelCalls

Methods

doEmbed

▸ doEmbed(options): Promise<EmbeddingModelV3Result>

Parameters

Name	Type
options	EmbeddingModelV3CallOptions

Returns

Promise<EmbeddingModelV3Result>

Implementation of

EmbeddingModelV3.doEmbed