Unified Context for Seamless Workflows

Anthropic has eliminated the friction of "rebriefing" AI agents by merging the memory systems of its standard chat interface and Claude Cowork. Previously, users often had to manually re-input project details when transitioning from a research or brainstorming phase in chat to an execution phase in Cowork. By unifying these systems, Claude now functions as a continuous assistant, automatically applying context—such as project parameters, conference details, or stakeholder information—across different modes of interaction.

Real-Time Memory Updates and User Control

Claude now updates its memory incrementally during active conversations rather than waiting for a session to conclude. This real-time synchronization ensures that information shared in a chat is immediately available to Cowork.

To address privacy and accuracy, Anthropic has introduced a transparency layer that allows users to:

• View: See exactly what information Claude has retained.
• Edit/Delete: Manually modify or remove specific memories.
• Sensitive Data Management: By default, Claude excludes sensitive personal information (e.g., political beliefs, health data, ethnicity). Users can opt-in to store these topics, though the system strictly prohibits saving high-risk data like Social Security numbers or government-issued IDs, regardless of user preference.

This update is currently rolling out across Free, Pro, and Max plans on web, desktop, and mobile platforms.