Claude’s Memory Is Now Shared Across Chats and Cowork, Thanks to Anthropic

Claude Memory System Update

According to НВ — Техно: Anthropic has rolled out a unified memory system for Claude that works across both chat conversations and Cowork. The assistant can now draw on information from earlier chats when handling tasks, and users are able to review, edit, and delete the stored data. Previously, content from regular conversations didn't carry over into Cowork, which made the feature much less practical. For people juggling quick questions and long-running projects, this removes a major friction point.

Claude can now add new items to memory while a conversation is in progress, rather than waiting until the exchange is over. The new capability is meant to make the assistant more responsive and better aligned with users' needs in real time. It comes enabled by default on Free, Pro, and Max plans, and is available on the web, desktop apps, and mobile apps for Android and iOS. For mobile users, updating Claude to the latest version is required to access these features.

Privacy and Data Safety

By default, Claude won't save personal or sensitive information-including health details, race, ethnicity, religious beliefs, political views, or gender identity. Users can choose to allow sensitive topics to be stored, and if such content does end up in memory, the app will flag it. Claude also won't retain data that violates its usage policies, such as government-issued ID numbers, Social Security numbers, criminal history, or immigration status.

TechCrunch covered the update, noting that these changes significantly improve how people interact with the assistant by making it more adaptive and easier to use across different scenarios. This shift also reflects a broader push in AI toward more intuitive, user-friendly systems. Giving people direct control over what Claude remembers helps them feel more confident about privacy and security, and it underscores Anthropic's focus on tailoring its technology to the needs of its customers.

This update is part of Anthropic's broader initiative to enhance AI interactions, similar to their recent approach of embedding an invisible code in Claude's text outputs. This innovation aims to ensure that AI-generated content can be easily identified, further improving user trust and transparency. To explore how these advancements are reshaping the landscape of AI, check out the details in our article on identifying AI outputs.