Top News

Anthropic has announced an update to the Claude memory system that makes information remembered in Claude Chat available in Claude Cowork, and vice versa. The company says the change is designed to reduce the need for users to repeatedly explain their projects, preferences, and working context when moving between Claude’s different experiences.

Previously, users could encounter a separation between the context available in Claude chat and Cowork. With the new system, Claude can use a shared memory across both environments, making it easier to move from a conversation to a task without starting from scratch.

Claude Memory Updates as Users Continue Conversations

Another important change is how Claude builds its memory. Anthropic says Claude can add information to memory as users chat, rather than waiting until a conversation has ended to create a summary. This allows memories to be updated quickly and shared between Chat and Cowork while work is still in progress.

For example, information discussed during a conversation can become part of Claude’s memory and then be available when the user later asks Cowork to work on a task. This could be particularly useful for longer projects that involve both conversational planning and task-based work.

Users Get More Control Over What Claude Remembers

Anthropic is also putting great emphasis on user control. Claude’s remembered information is organized into Topics under Settings > Memory, where users can see what AI has stored. Users can edit or delete individual memories, giving them more control over the information Claude retains.

Claude also does not automatically store topics considered sensitive, such as health or beliefs. Users can choose to enable the option to include sensitive topics in memory if they want Claude to remember such information.

Memory Availability Across Claude Plans

According to Anthropic’s release notes, memory is enabled by default for Free, Pro, and Max plans. For Team and Enterprise organizations, memory is off by default, allowing organizations to decide whether to enable the feature.

The unified memory system is available across Claude Chat and Cowork in the cloud, making the feature more consistent for users who regularly switch between the two experiences.

What The Update Means For Claude Cowork

The update is particularly significant for Claude Cowork, Anthropic’s task-oriented AI experience. Cowork can now begin a task with information Claude has already learned from previous interactions, instead of requiring users to provide the same background again.

This could make Cowork more useful for ongoing projects, where users may first discuss ideas with Claude in Chat and later ask Cowork to carry out work. Shared memory creates a more continuous experience between conversation and task execution.

Anthropic Focuses on Personalized AI

The latest update reflects the broader shift among AI assistants towards persistent, personalized interactions. Instead of treating every conversation a completely new session, Claude can now build a longer-term understanding of information users to choose to make available through memory.

At the same time, Anthropic’s controls are intended to ensure users can see and manage what Claude remembers. By combining cross-platform memory with editing and election controls, the company is positioning memory as a user-managed feature rather than an invisible background system.

Conclusion

Anthropic’s shared memory update makes Claude Chat and Claude Cowork more like a connected AI assistant. By allowing memories to move between the two experiences, Claude can retain useful context while users switch between conversation and task-based work. The ability to view, edit, and delete memories also gives users greater control over their stored information.

The update could make Claude more convenient for long-term projects, while marking another step towards AI assistants that can provide more personalized and continuous experiences.