memory

To persist context across AI sessions, use memory_store and memory_retrieve interactions to save and load decisions, preferences, and context. This maintains continuity by storing episodic, semantic, and procedural memories for future interactions.

Loading previous session context at startup requires retrieving stored memories via memory_retrieve. This restores previous decisions, project conventions, and long-term knowledge, allowing the agent to resume work smoothly without losing prior discoveries.

Yes, you can store different memory types including episodic, semantic, and procedural memories. By defining inputs such as content, memory_type, level, tags, and source, you manage durable, shared memory for project conventions and long-term knowledge.

To save agent decisions and discoveries as they happen, apply memory_store during ongoing work. This builds a usable history by persisting learnings and context with defined inputs like content and tags, ensuring accessibility across sessions.

Yes, this memory persistence approach works without external dependencies. It relies on defined inputs through memory_store and memory_retrieve to manage contextual memories, ensuring previous decisions and context are accessible across agent interactions.

You need episodic, semantic, and procedural memories when maintaining continuity across AI sessions. Use memory_type to categorize contextual data, applying memory at session start, ongoing work, and session end to persist decisions, preferences, and long-term knowledge.