Anthropic has updated Claude’s memory system so remembered context can move between standard Claude chats and Claude Cowork sessions, according to Engadget. The change means Claude can use relevant context from a user’s prior chats when Cowork is running a cloud task, and context from Cowork can also carry back into the broader Claude experience. Engadget describes the update as a shift from a more fragmented memory setup. Claude has had memory since last summer, but the publication says those memories have often been separated across experiences, sometimes intentionally and sometimes because of architectural limits. The new release is meant to make Claude’s memory more continuous across Anthropic’s chatbot and its agentic computer-use product. Cowork is Anthropic’s computer-use agent, which Engadget says the company released at the start of the year. It is designed to automate routine tasks users repeat during the day. Anthropic’s framing, as reported by Engadget, is that users should have to repeat themselves less often because Cowork can now draw on context Claude already has from prior conversations. The timing also connects to Anthropic’s Chrome work. Engadget reports that Anthropic recently upgraded the Claude for Chrome browser extension so Cowork is available through the plugin. Since that change, conversations in the Chrome side panel become part of the user’s usage history, according to the report. Anthropic is also adding more explicit memory controls. Engadget says users have recently been able to see Claude’s memories grouped by topic and edit the details of each topic from settings. The new addition is an “include sensitive topics in memory” option. By default, according to Anthropic as cited by Engadget, Claude does not store topics tied to sensitive personal areas such as health, race, ethnicity, religious beliefs, politics, gender identity and similar categories. If a user enables the sensitive-topics setting, Claude can begin adding those areas to memory, and Engadget reports that users will receive a notification when that happens. The toggle is not retroactive. Engadget reports that Anthropic will not have Claude add past conversations into memory after a user enables sensitive-topic storage. Users can turn the feature off, and can edit or delete saved memory entries. Anthropic also says Claude will still not save certain information, including Social Security numbers, criminal history and immigration status, even when the sensitive-topic setting is enabled. Claude memory is available and enabled by default for Free, Pro and Max accounts, according to Engadget. Mobile users need the latest Claude app on Android or iOS to see the new updates. Who benefits: Heavy Claude users who move between chat, Chrome and Cowork should benefit most because the assistant can retain more context across those surfaces. Teams evaluating agentic workflows may also find the continuity useful for repetitive cloud-run tasks. Who's exposed: Users who share sensitive context with Claude need to review the new memory controls carefully. Anthropic says sensitive-topic storage is off by default and remains editable or deletable, but the feature increases the importance of checking what Claude has saved.