Amazon DynamoDB now supports real-time vector search

This article announces the general availability of vector search for Amazon DynamoDB, enabling real-time indexing and searching of vectors at scale.

• Store vector embeddings alongside other attributes using models from Amazon Bedrock or custom models
• Achieve single-digit millisecond latency with 99%+ recall at any scale, even trillions of vectors
• Create vector indexes and run approximate nearest neighbor searches with attribute filtering
• Maintain serverless benefits: zero infrastructure management, zero downtime, no maintenance windows
• Enable use cases like semantic retrieval for AI agents, product similarity search, personalized advertising, RAG, and recommendation systems

DynamoDB vector search provides predictable performance for vector-based applications without operational overhead, available across all commercial AWS Regions and GovCloud.