What is Episodic memory?

Episodic memory is the experience-based half of an agent's long-term memory: a log of specific things that happened, this user asked for X last Tuesday, the deploy on Friday failed for reason Y, the previous attempt at this refactor was abandoned because Z. It is defined in contrast to semantic memory, which stores general, timeless facts (the API takes a token in this header) stripped of when or how they were learned. Episodic memories keep their context, the event, the time, the sequence, which is what lets an agent reason about its own history: avoid a path that failed before, resume where a prior session left off, or explain a decision by pointing to the moment it was made. The distinction is conceptual rather than a hard storage boundary; many memory systems capture episodes and later distill durable facts from them. For agents, good episodic memory turns a string of independent sessions into a continuous working relationship, and the retrieval problem, surfacing the right past episode for the current situation, is as important as recording the episode in the first place.