Announcing vector search for Amazon ElastiCache

AWS has announced vector search for Amazon ElastiCache, a new capability that enables indexing, searching, and updating billions of high-dimensional vector embeddings with microsecond latency and up to 99% recall.

• Supports semantic caching for generative AI applications to reduce cost and latency
• Enables conversational memory and retrieval-augmented generation (RAG)
• Provides low-latency vector search with HNSW and FLAT algorithms
• Supports hybrid queries with vector and non-vector attributes
• Available for ElastiCache version 8.2 for Valkey in all AWS Regions

The feature allows developers to build more efficient and context-aware generative AI applications by leveraging in-memory vector storage and fast similarity search capabilities.