Redis Vector Search

Redis Stack includes vector similarity search capabilities through the RediSearch module, enabling Redis to serve as a vector database alongside its traditional caching and data structure roles. This allows teams to add AI features without introducing new infrastructure.

Features include: FLAT and HNSW indexing algorithms, hybrid search combining vectors with tag/text/numeric filters, real-time index updates, and the familiar Redis performance characteristics. Vector search integrates with Redis's other data structures and caching capabilities.

For AI engineers already using Redis, adding vector capabilities is attractive: no new systems to operate, sub-millisecond latencies, and combining caching with semantic search. Use cases include: session-aware semantic search, real-time recommendation, and RAG applications needing fast retrieval. Consider dedicated vector databases for very large scale or advanced features not available in Redis.