Claude Cowork Now Remembers Your Chat History

Anthropic has rolled out a change to Claude’s memory system that ends the frustrating loop of re-explaining details to its AI agent. The company announced Tuesday that chat and Claude Cowork now share one unified memory, so anything Claude learns in one mode is automatically available in the other.

This means users can move from planning in chat to executing tasks in Cowork without starting from scratch. Previously, you might have spent time building project ideas in conversation, only to have to repeat all those specifics when you switched to action mode. Now, the AI behaves like a single continuous assistant rather than two separate products.

Anthropic is also making memory visible and editable, letting users read, update, or delete any stored topic. As you chat, Claude adds new memories on the fly instead of waiting for a conversation summary, so information syncs faster between the two experiences, even mid-conversation.

By default, Claude will not save sensitive personal data such as health information, race, ethnicity, religion, politics, or gender identity. Users can opt in to storing these by enabling a toggle called “include sensitive topics in memory.” The app will also alert you whenever it saves a sensitive topic.

Certain items are never stored, including government-issued IDs, Social Security numbers, criminal history, and immigration status, as well as anything violating its acceptable use policy. The memory feature is turned on by default for Free, Pro, and Max plans across web, desktop, and mobile. iOS and Android users need to update their apps to the latest version to access it.