Voyage AI

service active paid

A specialized embedding model provider focused on retrieval-optimized embeddings. Voyage AI offers embedding models trained specifically for search and RAG use cases, with models optimized for different domains (code, legal, finance) and multilingual support.

Implements

Concepts this tool claims to implement:

  • Embedding primary

    Embedding API with models like voyage-large-2, voyage-code-2, voyage-law-2. Optimized for retrieval tasks with high performance on MTEB benchmarks. Up to 16K token context.

  • Retriever secondary

    Reranking API for improving retrieval quality. Cross-encoder reranking to refine initial retrieval results.

Integration Surfaces

  • REST API
  • Python SDK
  • LangChain integration
  • LlamaIndex integration

Details

Vendor
Voyage AI
License
Proprietary
Runs On
cloud
Used By
human, agent, system

Notes

Voyage AI focuses exclusively on embeddings, unlike general-purpose providers. Their domain-specific models (code, legal) often outperform general embeddings for those use cases. Good choice for teams building RAG systems that need high retrieval quality.