Anthropic makes Claude memory work across chat and Cowork

Anthropic is making Claude’s memory shared between regular chat and Claude Cowork, so context learned in one surface can carry into the other instead of forcing users to brief the assistant again.

The change turns memory into a more continuous layer across Claude products. TechCrunch and CNET report that Claude can now update memory topics while a conversation is still happening, not only after a chat ends. Users can open memory settings to see saved topics, edit them, delete them, pause memory or reset it.

The privacy controls matter because memory is becoming more useful and more sensitive at the same time. Sensitive topics are off by default, but users can opt in to saving some of them. Anthropic says Claude still will not save certain categories such as government ID numbers, financial account numbers, criminal history or immigration status, even when sensitive-topic memory is enabled.

For everyday users, the practical effect is less repeated context when moving from planning to execution. A project discussed in chat can be picked up by Cowork with names, preferences and details already available, while still giving the user a visible place to clean up what Claude remembers.