embedMany()

Embed several values using an embedding model. The type of the value is defined by the embedding model.

embedMany automatically splits large requests into smaller chunks when the model has a limit on either the number of embeddings or the UTF-8 input bytes that can be processed in a single call. Providers can use a conservative byte budget to keep requests below aggregate token limits without adding a tokenizer to the AI SDK core package. An individual value larger than the byte budget is sent in its own call because splitting it would change the resulting embedding.

1
import { embedMany } from 'ai';
2

3
const { embeddings } = await embedMany({
4
  model: 'openai/text-embedding-3-small',
5
  values: [
6
    'sunny day at the beach',
7
    'rainy afternoon in the city',
8
    'snowy night in the mountains',
9
  ],
10
});

Import

import { embedMany } from "ai"

API Signature

Parameters

Returns