Artificial intelligence startup Anthropic announced on August 25 that it has made a major upgrade to the memory system of its AI assistant Claude, officially breaking down the memory barriers between the chat system (Chat) and Claude Cowork, enabling seamless sharing and real-time synchronization of cross-platform context information.
The core of this update addresses the pain point of information gaps in complex workflows for AI assistants. Previously, project ideas or details formed during the Claude chat phase often needed to be re-entered when transitioning to the Cowork execution phase.

After the system upgrade, Claude can now record key topics in real time during the chat process, without waiting for the conversation to end for summarization. When users generate meeting agendas or handle specific projects, Cowork can automatically retrieve key elements such as people, locations, and topics discussed previously, significantly improving collaborative efficiency from research exploration to implementation.
In terms of data privacy and user control, Anthropic has opened up transparent management of memory information, allowing users to view, edit, or delete specific records at any time. The system will not default to storing sensitive personal information such as health conditions, political views, or religious beliefs, but provides optional switches and prompts users when sensitive records are triggered; it also strictly prohibits saving sensitive documents such as ID numbers and social security numbers, as well as illegal and non-compliant content.

Currently, this unified memory mechanism is enabled by default on the web version, desktop, and mobile applications for free, professional (Pro), and premium (Premium) users. As the focus of competition in generative AI shifts from single dialogue capabilities to deep workflow integration, Anthropic's cross-product memory integration further strengthens Claude's ecological competitiveness as a continuous intelligent collaboration assistant.
Join Now