DynamoDB learned vector search so you can fire your vector database

Amazon DynamoDB now supports native vector search, letting you store vector embeddings alongside operational data and run similarity queries without replicating to a separate vector store. It delivers single-digit millisecond latency at 99%+ recall, scales to trillions of vectors, and supports up to 4096 dimensions with Euclidean, Cosine, and Dot product distance functions plus inline filtering. The feature uses DynamoDB's existing serverless pay-per-request model, so there's no infrastructure to provision or maintain. This is particularly useful for RAG, recommendation engines, and anomaly detection workloads where your data already lives in DynamoDB.