Anthropic has released Memory on Claude Managed Agents to public beta. This feature enables agents to learn from past sessions and share knowledge with other agents. Memories are stored as files on a filesystem, leveraging the same bash and code execution tools already used in Managed Agents. Multiple agents can access the same memory store concurrently without overwriting each other. Scoped permissions allow enterprise-level read-only stores or per-user read-write access. Every memory write is logged as a session event in the Claude console, giving teams full traceability with rollback and redaction capabilities. Memories can also be exported and managed via API.
Sort: