Anthropic has unified Claude's memory across chat and Cowork so context carries between both, updating live during long Cowork sessions instead of only after chats end. By default Claude won't store sensitive data like health, religion, race, ethnicity, or gender identity, though users can opt in via settings; some data (SSNs, government IDs, criminal history, immigration status) is never stored. Memory is on by default for Free, Pro, and Max plans but off by default for Enterprise and Teams.
Questions this post answers
Does Claude's memory now work the same way in chat and Cowork?
Yes, Anthropic unified Claude's memory so it bridges both chat and Cowork rather than keeping Cowork's memory tied only to individual projects. Context built up in chat, such as ongoing project status, is available the moment you hand Cowork a task, and anything Cowork learns carries back to chat. Track how Claude's shared memory changes cross-tool workflows by following updates on daily.dev.
What personal information does Claude refuse to store in memory by default?
By default, Claude won't store sensitive information such as health data, religious beliefs, race, ethnicity, or gender identity, though users can opt in through settings to include these topics, with no option to select only certain ones. Some data is never stored regardless of settings: SSNs, government ID numbers, criminal history, and immigration status. Developers evaluating AI privacy controls can follow these policy details on daily.dev.
Is Claude's new memory feature enabled by default for all account types?
Unified memory is on by default for Free, Pro, and Max plans across web, desktop, and mobile, but for Enterprise and Teams plans all memory features are off by default. Even where enabled, the system withholds sensitive information from memory unless a user manually opts in. Teams weighing Claude plan settings can keep up with rollout details on daily.dev.