Redis

Redis stores data in memory, which makes reading and writing it extremely fast. Developers use it as a cache that sits in front of a slower database, as a session store for logged-in users, as a message queue, and for leaderboards, rate limiting, and counters. Its data types, such as lists, sets, sorted sets, hashes, and streams, make these patterns easy to build with simple commands. It’s one of the most widely deployed pieces of infrastructure on the web.

For AI, Redis has added vector search, so applications can store embeddings and find similar items for retrieval-augmented generation, recommendations, and semantic search. Semantic caching stores answers to earlier questions and returns them for similar new ones, which reduces cost and latency when calling language models. Redis also holds conversation memory for agents. It runs as open-source software, as Redis Cloud, and in enterprise deployments.

Licensing has been a topic of debate: Redis changed its license in a way that led to a community fork called Valkey, backed by cloud providers, and later offered an open-source license option again. Memory is more expensive than disk, so large datasets can cost a lot, and durability needs configuration. Alternatives include Valkey, Memcached for simple caching, DragonflyDB, and cloud services like Amazon ElastiCache. Redis suits applications where speed is critical.