Anthropic expands Claude's memory to span chat and Cowork sessions
Anthropic updated Claude so its memory now works across both the chatbot and the Cowork automation tool, and added new user controls for sensitive topics.
Anthropic announced that Claude's memory capability has been unified across its chat interface and the Cowork computer-use assistant, allowing the model to carry over context between the two environments. This change means users will not have to restate information when switching from a conversation to an automated task. The company also introduced a settings panel where memories are organized by topic, and added an optional "include sensitive topics in memory" switch for subjects such as health, race, religion, politics, and gender identity.
When enabled, each addition triggers a notification, yet Claude still excludes highly personal identifiers like social security numbers or criminal history. The memory feature remains on by default for Free, Pro, and Max plans, and the update is available on the newest Android and iOS Claude apps.
Why it matters
The upgrade lets users interact more seamlessly with Claude and gives them control over what personal data the AI retains.
In this story
