ZeroEntropy Provider

zeroentropy-ai/zeroentropy-ai-provider is a community provider that uses ZeroEntropy to provide text embedding and reranking support for the AI SDK.

Setup

The ZeroEntropy provider is available in the zeroentropy-ai-provider module. You can install it with:

pnpm add zeroentropy-ai-provider

Provider Instance

You can import the default provider instance zeroentropy from zeroentropy-ai-provider:

1
import { zeroentropy } from 'zeroentropy-ai-provider';

If you need a customized setup, you can import createZeroEntropy from zeroentropy-ai-provider and create a provider instance with your settings:

1
import { createZeroEntropy } from 'zeroentropy-ai-provider';
2

3
const zeroentropy = createZeroEntropy({
4
  apiKey: process.env.ZEROENTROPY_API_KEY ?? '',
5
});

You can use the following optional settings to customize the ZeroEntropy provider instance:

• baseURL string

  The base URL of the ZeroEntropy API. The default prefix is https://api.zeroentropy.dev/v1.

• apiKey string

  API key that is being sent using the Authorization header. It defaults to the ZEROENTROPY_API_KEY environment variable. Obtain your API key from the ZeroEntropy Dashboard.

• headers Record<string,string>

  Custom headers to include in the requests.

• fetch (input: RequestInfo, init?: RequestInit) => Promise<Response>

  Custom fetch implementation. Defaults to the global fetch function. You can use it as a middleware to intercept requests, or to provide a custom fetch implementation for e.g. testing.

Text Embedding Models

You can create models that call the ZeroEntropy embeddings API using the .textEmbeddingModel() factory method.

1
import { zeroentropy } from 'zeroentropy-ai-provider';
2

3
const embeddingModel = zeroentropy.textEmbeddingModel('zembed-1');

You can use ZeroEntropy embedding models to generate embeddings with the embed or embedMany function:

1
import { zeroentropy } from 'zeroentropy-ai-provider';
2
import { embed, embedMany } from 'ai';
3

4
// Single embedding
5
const { embedding } = await embed({
6
  model: zeroentropy.textEmbeddingModel('zembed-1'),
7
  value: 'sunny day at the beach',
8
});
9

10
// Batch embeddings
11
const { embeddings } = await embedMany({
12
  model: zeroentropy.textEmbeddingModel('zembed-1'),
13
  values: ['first document', 'second document'],
14
});

ZeroEntropy embedding models support additional provider options that can be passed via providerOptions.zeroentropy:

1
import { zeroentropy } from 'zeroentropy-ai-provider';
2
import { embed } from 'ai';
3

4
const { embedding } = await embed({
5
  model: zeroentropy.textEmbeddingModel('zembed-1'),
6
  value: 'sunny day at the beach',
7
  providerOptions: {
8
    zeroentropy: {
9
      inputType: 'document',
10
      dimensions: 1280,
11
    },
12
  },
13
});

The following provider options are available:

• inputType 'query' | 'document'

  Whether the input is a search query or a document to index. Use 'query' for retrieval queries and 'document' for content being indexed. Defaults to 'document'.

• dimensions 2560 | 1280 | 640 | 320 | 160 | 80 | 40

  Output vector dimensions. Defaults to 2560. Supports Matryoshka representations for flexible dimensionality reduction.

• latency 'fast' | 'slow'

  'fast' for sub-second latency; 'slow' for higher throughput limits.

Model Capabilities

Reranking Models

You can create models that call the ZeroEntropy reranking API using the .rerankingModel() factory method.

1
import { zeroentropy } from 'zeroentropy-ai-provider';
2
import { rerank } from 'ai';
3

4
const result = await rerank({
5
  model: zeroentropy.rerankingModel('zerank-2'),
6
  query: 'talk about rain',
7
  documents: ['sunny day at the beach', 'rainy day in the city'],
8
  topN: 2,
9
});

ZeroEntropy reranking models support additional provider options that can be passed via providerOptions.zeroentropy:

1
import { zeroentropy } from 'zeroentropy-ai-provider';
2
import { rerank } from 'ai';
3

4
const result = await rerank({
5
  model: zeroentropy.rerankingModel('zerank-2'),
6
  query: 'talk about rain',
7
  documents: ['sunny day at the beach', 'rainy day in the city'],
8
  topN: 2,
9
  providerOptions: {
10
    zeroentropy: {
11
      latency: 'fast',
12
    },
13
  },
14
});

The following provider option is available:

• latency 'fast' | 'slow'

  'fast' for sub-second latency; 'slow' for higher throughput limits.

Model Capabilities