Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/patelvivekdev/voyageai-ai-provider/llms.txt

Use this file to discover all available pages before exploring further.

The VoyageRerankingOptions type defines the configuration options you can pass when reranking documents using Voyage AI’s reranking models.

​
Properties

​
returnDocuments
boolean
Whether to return the documents in the response.Default: false
  • If false: The API returns a list of {"index", "relevance_score"} where “index” refers to the index of a document within the input list.
  • If true: The API returns a list of {"index", "document", "relevance_score"} where “document” is the corresponding document from the input list.
​
truncation
boolean
Whether to truncate the input to satisfy the “context length limit” on the query and the documents.Default: true
  • If true: The query and documents are truncated to fit within the context length limit before being processed by the reranker model.
  • If false: An error is raised when:
    • The query exceeds the token limit:
      • 8,000 tokens for rerank-2.5 and rerank-2.5-lite
      • 4,000 tokens for rerank-2
      • 2,000 tokens for rerank-2-lite and rerank-1
      • 1,000 tokens for rerank-lite-1
    • The sum of tokens in the query and any single document exceeds:
      • 32,000 for rerank-2.5 and rerank-2.5-lite
      • 16,000 for rerank-2
      • 8,000 for rerank-2-lite and rerank-1
      • 4,000 for rerank-lite-1

​
Usage

​
Basic reranking

import { voyage } from '@voyageai/ai-provider';

const result = await voyage.reranking('rerank-2.5').doRerank({
  query: 'What is machine learning?',
  documents: [
    'Machine learning is a subset of artificial intelligence.',
    'Python is a programming language.',
    'Neural networks are used in deep learning.',
  ],
});

// result.rankings contains [{ index: number, relevanceScore: number }, ...]

​
Reranking with document return

import { voyage } from '@voyageai/ai-provider';

const result = await voyage.reranking('rerank-2.5').doRerank({
  query: 'What is machine learning?',
  documents: [
    'Machine learning is a subset of artificial intelligence.',
    'Python is a programming language.',
    'Neural networks are used in deep learning.',
  ],
  returnDocuments: true,
});

// result.rankings now contains:
// [{ index: number, document: string, relevanceScore: number }, ...]

​
Reranking without truncation

import { voyage } from '@voyageai/ai-provider';

const result = await voyage.reranking('rerank-2.5').doRerank({
  query: 'What is machine learning?',
  documents: [
    'Machine learning is a subset of artificial intelligence.',
    'Python is a programming language.',
  ],
  truncation: false,
});

// Throws an error if query or documents exceed token limits

​
Semantic search with reranking

import { voyage } from '@voyageai/ai-provider';
import { embed } from 'ai';

const documents = [
  'Machine learning is a subset of artificial intelligence.',
  'Python is a programming language.',
  'Neural networks are used in deep learning.',
  'Data science involves analyzing data.',
  'Cloud computing provides on-demand resources.',
];

const query = 'artificial intelligence and machine learning';

// First, get initial candidates (e.g., from a vector database)
// Then rerank them for better relevance
const result = await voyage.reranking('rerank-2.5').doRerank({
  query,
  documents,
  returnDocuments: true,
});

// Get top 3 most relevant documents
const topDocuments = result.rankings
  .slice(0, 3)
  .map((ranking: any) => ranking.document);

console.log('Top documents:', topDocuments);

​
Using different reranking models

import { voyage } from '@voyageai/ai-provider';

// Use the lite model for faster, lower-cost reranking
const litResult = await voyage.reranking('rerank-2.5-lite').doRerank({
  query: 'What is AI?',
  documents: ['Document 1', 'Document 2', 'Document 3'],
});

// Use the standard model for higher quality
const standardResult = await voyage.reranking('rerank-2.5').doRerank({
  query: 'What is AI?',
  documents: ['Document 1', 'Document 2', 'Document 3'],
});

Build docs developers (and LLMs) love

Get started for freeTalk to us