Databricks launches Agent Memory in beta for Agent Bricks agents
AIDatabricks introduces Agent Memory in beta on Agent Bricks, its agent developer platform, to store facts, preferences, and workflows for agents. Running on Lakebase, the memory uses managed sessions, on-demand memory tools, and a background pipeline that deduplicates and updates stored facts.
Source details
Databricks Blog · databricks.comPublished · added here
Original title: Introducing Agent Memory on Databricks