Anthropic unifies memory across Claude Cowork and chat so context persist across workflows

Anthropic has unified the memory system for Claude’s chat and Cowork, allowing user information to persist seamlessly across both workflows. This means when you interact with Claude, whether in standard chat or via Cowork for collaborative tasks, the assistant now builds on what it already knows about your context and priorities.

Claude maintains a live record of topics as you chat, rather than waiting until the conversation ends to summarize information. This update helps users avoid repetitive explanations and ensures that any context developed in one area, such as project deadlines or quarterly goals, is instantly accessible in both chat and Cowork. When Cowork runs tasks in the cloud, it uses chat memory, and any updates there flow back to chat conversations.

To enhance transparency and control, everything that Claude remembers is visible in a files-and-topics list within memory settings, where users may review, edit, or delete any entry. Users can pause or reset memory at any time. While sensitive topics such as health or religious beliefs are not stored by default, users can enable memory for these areas if needed. However, certain sensitive details, including identification numbers or immigration status, are never stored, and users are informed if memory cannot be updated for these reasons. This unified memory is active by default for users on Free, Pro, and Max plans across all platforms.