Voyage AI

What it does

Voyage AI provides embedding and reranking models behind API endpoints. Embedding models convert documents, images, and other unstructured data into dense numerical vectors that capture meaning, while rerankers output relevance scores between a query and multiple documents. Developers call the models through a standalone API or use automated embeddings in MongoDB Atlas Vector Search.

A typical workflow is semantic search and retrieval-augmented generation: embed a corpus, store the vectors, retrieve candidates, then rerank them before passing results to a language model. The standalone API is described as database-agnostic and works with any vector-capable database or AI framework, while the Atlas integration keeps embedding and search in one platform. Model families include text, multimodal, and domain-specific variants for legal, finance, and code.

The service suits developers and enterprise teams comfortable managing API keys and usage billing. Access is through the Voyage dashboard and API, or through Atlas. Pricing is consumption-based per token, with free token allowances per account; Batch API usage does not draw from those allowances.

What it can do

• Generate text embeddings for documents and queries using Voyage embedding models
• Generate multimodal embeddings from text, images, and video frames
• Rerank candidate documents against a query to refine retrieval results
• Process large batches of requests through a Batch API with a 12-hour completion window
• Use automated embeddings in MongoDB Atlas Vector Search indexes

Works with

• MongoDB Atlas Vector Search (automated embeddings)

Before you choose

• Pricing is usage-based per token, billed to the account linked to the API key.
• Batch API usage does not deduct from free token balances and is billed at a 33% discount.
• Several model rows in the pricing table lack explicit prices in the source; confirm current rates on the pricing page.

Listed under

• Data

Get started

Sign in to the Voyage dashboard or follow the quickstart tutorial.

1. Sign in to the Voyage AI dashboard and create an API key.
2. Install the voyageai Python client and set the VOYAGE_API_KEY environment variable.
3. Call the embed endpoint with a chosen model and input_type, then store or index the resulting vectors.

Billed per thousand or million tokens; free allowances per account do not apply to Batch API. Multimodal billed per token and per billion pixels. Rerankers billed on processed tokens. Verify rates on the Voyage pricing page.

Source: docs.voyageai.com/docs/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Introduction ↗docs.voyageai.com
• Pricing ↗docs.voyageai.com
• Voyage AI | Home ↗www.voyageai.com
• Voyage AI : Embeddings and Rerankers | MongoDB ↗www.mongodb.com
• Pricing | MongoDB ↗www.mongodb.com
• Welcome to the MongoDB Docs - MongoDB Documentation - MongoDB Docs ↗www.mongodb.com

Company website

mongodb.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

manual_review

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.