Skip to main content

Public Preview: DiskANN vector indexing and search in Azure Cosmos DB NoSQL

Published date: May 21, 2024

Azure Cosmos DB for NoSQL built-in vector search is now in public preview. This feature can be used with a flat vector index, quantized vectors, and with a vector index powered by DiskANN, a state-of-the-art suite of vector indexing algorithms developed at Microsoft Research. This enables your applications to perform a vector similarity search with high performance,accuracy, and scale.  
You can easily combine vector search with Azure Cosmos DB's rich query filters. Plus, take advantage of Azure Cosmos DB's flexible schema-less document structures, serverless or provisioned throughout models, 99.999% availability, instant autoscale, and more! 
Simplify your RAG and Generative AI architectures and keep your data and vectors in sync by leveraging Azure Cosmos DB as your one database for transactional and vector data workloads.  

Learn more

  • Azure Cosmos DB
  • Features
  • Microsoft Build

Related Products