Anthropic adds memory to Claude Managed Agents

TL;DR AI
2 min readKey summary
Anthropic has added a public beta memory layer to Claude Managed Agents.
The memory is stored as files, can be shared across agents without overwriting, and works with the file system and CLI.
Developers can manage, export, restore, or delete memory through the API and Claude Console.
The update gives Claude agents more persistent context for long-running workflows while keeping memory transparent and permissioned.



