Ultra-fast in-memory store used for agent caching, sessions, and vectors.
Category: Memory & Data · Type: In-memory data store + vector search · Website: https://redis.io
Overview
Redis is an in-memory data store prized for speed. For agents it serves as a cache, session/short-term memory store, message broker, and — via Redis vector search — a low-latency vector database.
Key Features
- Sub-millisecond in-memory access
- Vector search for embeddings
- Caching, sessions, queues, pub/sub
- Managed and self-hosted options
Role in the Agentic AI Stack
The speed layer for agents — short-term memory, caching, and fast vector lookups — alongside durable stores like MongoDB.
Part of Agentic AI Landscape → Memory & Data