Back to feed
News Story
Claude (X)
1 sources

Claude now has unified memory across chat and Cowork, with user control

Anthropic announced that Claude now has a unified memory system shared between chat and Claude Cowork, allowing users to view, edit, or delete memory topics in Settings. The feature is enabled by default on Free, Pro, and Max plans, and users can save specific information by saying "remember this."

SynthePulse Insight · AI deep reading

Claude Unified Memory: Persistent Context Across Chats and Cowork, Under Your Control

Version 1 · 1 source

Anthropic introduces Claude's unified memory, spanning chats and Claude Cowork, allowing users to view, edit, or delete memory topics in settings, with sensitive topics not recorded by default.

  • Claude now shares the same memory between chats and Claude Cowork, allowing tasks to draw context from previous conversations.
  • All memories are stored as a list of topics in settings, where users can read, edit, or delete each one.
  • Memory updates automatically with chats, and can also be explicitly saved by saying "remember this."
  • Sensitive topics like health and religious beliefs are not saved to memory by default, unless users enable this in settings.
  • Memory is enabled by default on Free, Pro, and Max plans, and users can view it anytime in settings.
Open section navigationUnified Memory: Shared Context Across Products

Unified Memory: Shared Context Across Products

On August 25, 2026, Anthropic announced that Claude now has unified memory across chats and Claude Cowork. According to the official X post, when users delegate tasks to Cowork, it starts with information Claude already knows from chats, such as discussed projects, manager preferences, or last quarter's clients.

The core of this feature is eliminating context fragmentation across products, enabling Claude to carry over conversation history across different interfaces. The official description emphasizes "one memory," indicating this is an integration of existing memory systems rather than a new standalone memory.

User Control: Transparent Memory Management

The official post details how memory is managed: everything Claude remembers is stored as a list of topics in settings, where users can read, edit, or delete each one. This design clearly puts control in the user's hands, aligning with the promise that "you decide what's inside."

Memory not only updates automatically with chats, saving new details, but also supports explicit saving by saying "remember this." This provides both automatic and manual paths for memory writing, balancing convenience and precision.

Privacy Protection: Default Exclusion of Sensitive Topics

For potentially sensitive topics like health and religious beliefs, the official statement says these are not saved to memory by default, unless users actively enable this in settings. This mechanism aims to reduce privacy risks, giving users the final say on storing sensitive information.

Memory is enabled by default on Free, Pro, and Max plans, and users can view it anytime in the "Memory" section of settings. The official also provides a detailed explanation page to guide users through the feature's specifics.

Credibility boundary

This report is based on content from Anthropic's official X account, a first-party source. All feature descriptions come from official statements but have not been independently verified; actual experience may vary by version or region.

Insight takeaway

Claude's unified memory bridges context between chats and Cowork, while emphasizing user control and privacy through editable topic lists and default exclusion of sensitive topics.

Primary report

Claude (X)

Primary source