Agent Memory
Agent memory is what an AI agent keeps and reuses beyond a single turn — working memory in the context window, plus longer-term stores it can write to and retrieve from later. It's how an agent stops starting every session from zero.
A bare model forgets everything the moment a task ends. Agent memory is the layer that fixes that. The short-term part is working memory — what’s in the context window right now. The long-term part lives in external stores (vector or graph databases) the agent can write to and pull back later by relevance or ID.
Practitioners split long-term memory into kinds: semantic (facts and concepts), episodic (timestamped past interactions), and procedural (learned skills and routines). Once you move information out to a store and give the agent a way to retrieve it, you own retrieval speed, decay, and organization — which is why memory is increasingly treated as its own engineering discipline rather than a side effect of the prompt.
