Enables conversation memory for LLMs by storing chat history and retrieving relevant memories via embedding-based semantic search, supporting tools like store_turn and flashback_memory.
