lakebase-memory-patterns

What problem does it solve?

This Skill enables stateful AI agents to maintain conversation continuity and remember user preferences across sessions by leveraging Databricks Lakebase for both short-term and long-term memory.

Core Features & Use Cases

• Short-Term Memory: Ensures conversation continuity within a single session using CheckpointSaver.
• Long-Term Memory: Stores user preferences and insights across multiple sessions using DatabricksStore with vector embeddings.
• Graceful Degradation: Allows the agent to function even if memory tables are unavailable.
• Thread ID Resolution: Intelligently determines the conversation thread ID for state management.
• Use Case: An AI customer support agent can remember a user's previous issues and preferences, providing a more personalized and efficient support experience without starting from scratch each time.

Quick Start

Initialize the short-term memory with your Lakebase instance name and use its checkpointer when compiling your LangGraph workflow.