Redis

A linear, one-concept-per-lesson path through Redis, the in-memory store agents lean on for fast, ephemeral state, from your first client connection to a multi-agent system using Redis for memory, queue, and cache. 24 lessons · 3 tiers.

Prerequisites: Completion of the pgvector course, which covers durable, queryable long-term storage for embeddings. This course is the other half of that picture: Redis holds an agent's fast, ephemeral working memory (recent messages, a cached answer, an in-flight rate-limit counter), while pgvector holds what needs to survive and be searched by meaning over the long run. Lesson 12 draws that line explicitly.

No AI model calls in this course, it's pure Redis and Python (plus LangGraph in Lesson 14, with no LLM node in its graph). Lessons that talk about "an LLM response" or "an agent" simulate the expensive call with a deliberately slow local function, so nothing here needs a GOOGLE_API_KEY. Setup also needs a local Redis, run via `docker compose up -d redis` from the project root, the redis-stack-server image (not plain redis), exposed on localhost:6379, since the advanced lessons on vector and JSON search need the RedisJSON and RediSearch modules it ships with.

Built on Redis, the in-memory data store this course teaches.