Anthropic has rolled out a major update to Claude’s memory architecture, bridging the gap between its standard chat interface and the Claude Cowork environment. Previously, users frequently faced the frustrating hurdle of having to rebrief the AI whenever they transitioned from brainstorming ideas in chat to executing tasks in Cowork, as the two systems operated in silos.
With this update, Claude functions as a single, continuous assistant. Memories are now captured dynamically as conversations unfold rather than just at the end, allowing for a seamless shift from research to action. Furthermore, users gain complete visibility and control, with the ability to view, edit, or delete any stored topic at their discretion.
The feature is enabled by default across Free, Pro, and Max tiers on web, desktop, and mobile applications. To safeguard privacy, Anthropic has set default restrictions preventing the storage of highly sensitive personal data like government IDs and health records, while allowing user-controlled toggles for specific preference-based topics.
- Unified memory system linking chat and Claude Cowork
- Full user control to view, edit, or delete stored memory items
- Real-time memory updates occurring during ongoing conversations
- Default rollout across free and paid plans on web and mobile
Sources:
