Auto-Memory overview

Auto-Memory as a cross-cutting topic — agent-driven automatic memory extraction and persistence without explicit user action. Covers the mechanisms by which agents identify salient facts, preferences, and decisions during conversations and persist them for future sessions. Includes Claude Code's extract-memories fork that runs at turn-end, Mem0's LLM-driven extraction pipeline, Cline's memory bank updates, and the fundamental design tension between extraction accuracy (avoiding noise) and completeness (not missing important context). Auto-memory transforms ephemeral conversations into durable organizational knowledge automatically, but the quality of extraction determines whether the accumulated memories help or hinder future sessions.