
Amazon OpenSearch Serverless now offers a vector engine in preview that enables developers to build ML-augmented search and generative AI applications by querying billions of vector embeddings without managing infrastructure.


<div><p>This article announces the preview availability of a vector engine for Amazon OpenSearch Serverless, enabling developers to build ML-augmented search and generative AI applications without managing vector database infrastructure.</p><ul><li>Vector engine enables hybrid search combining vector embeddings with text-based keywords</li><li>Generates contextually relevant responses by querying billions of vector embeddings in milliseconds</li><li>Supports near real-time vector embedding updates without re-indexing or performance impact</li><li>Compatible with OpenSearch clients and open-source tools like LangChain</li><li>Powered by k-nearest neighbor (k-NN) search from open-source OpenSearch Project</li><li>Available in preview across eight AWS regions globally</li></ul><p>The vector engine provides a serverless, cost-efficient solution for building scalable generative AI applications without infrastructure management overhead.</p></div>


Related articles