Voyage AI site

Voyage AI develops embedding models and reranking models used to power search and retrieval over unstructured data. Its models are positioned as components of retrieval-augmented generation (RAG) pipelines, in which unstructured data is converted into embeddings, stored in a vector database, retrieved and reordered by a reranker, and passed as context to a large language model to produce responses with lower cost.

The product line spans general-purpose models intended to work across use cases and languages out of the box, domain-specific models optimized for industry data such as finance, legal and code, and company-specific models fine-tuned on a customer's own data and terminology. Announced releases include the voyage-3.5 and voyage-3.5-lite embedding models, the Voyage 4 model series, voyage-multimodal-3.5, voyage-context-3 (which combines chunk-level detail with global document context), the rerank-2.5 and rerank-2.5-lite rerankers, and a Batch API for large-scale workloads.

Business model

Voyage AI offers its embedding and reranker models to developers and enterprises, including self-serve access for general-purpose and domain-specific models and a sales-led path for company-specific fine-tuned models.