Anthropic released an update to Claude that unifies chat and Cowork memory, enabling shared memory across both and adding controls to view, edit, or delete stored topics
Feature update ● Confirmed 90% confidence first seen
Anthropic updated Claude’s memory so that information remembered in chat can persist and be used during Claude Cowork, rather than being limited to a single interaction stream. The update is enabled by default on Free, Pro, and Max plans and includes topic-by-topic controls to view, edit, or delete stored memory, while sensitive topics remain excluded by default unless users opt in.
Decision brief
- What changed
- Anthropic updated Claude so memory is now unified across chat and Claude Cowork, allowing retained context to persist between both experiences instead of staying within a single interaction stream or project. The update is enabled by default for Free, Pro, and Max plans and adds controls for users to view, edit, or delete stored memory topics, while sensitive topics remain excluded by default unless users opt in.
- Why it matters
- For leaders evaluating AI assistants for repeated knowledge work, this changes Claude from a more session-bound tool into one that can carry operational context across workflows, which can reduce repeated prompting and improve continuity in team use cases. The added user controls and default exclusion of sensitive topics also make memory more governable, which matters for balancing productivity gains against data-handling expectations. Decision-makers should care because persistent cross-workflow memory can increase usefulness, but default-on retention may require clearer internal usage guidance and data governance review.
- Evidence
- All three cited outlets report the same core change: The New Stack says Anthropic unified memory between chat and Cowork and enabled it by default on Free, Pro, and Max; SiliconANGLE and TechCrunch independently corroborate cross-experience persistence plus controls to view, edit, or delete memory and default exclusion of sensitive topics unless users opt in. The consistency across three separate reports supports the basic facts of the feature update.
- What remains uncertain
- The coverage does not establish how admin controls, enterprise policy settings, retention periods, or auditability work beyond user-level topic controls, so business-readiness for regulated environments is not yet clear. It also does not verify how reliably Claude distinguishes sensitive from non-sensitive information in practice, which is an important assumption if teams plan to rely on default protections.
- Monitor next
- Watch for Anthropic to publish enterprise/admin governance details for unified memory, especially policy controls, retention behavior, and audit features.
Analytical support, not advice — assumptions and open questions stated above.