Claude's Memory Now Works Across Chat and Cowork

Claude memory now works across Claude chat and Cowork, according to Anthropic. Context saved from one surface can help Claude continue work on the other, while users can inspect and control the topics Claude keeps. (Claude on X; Anthropic)

What the update changes

Memory is designed to make Claude more useful across longer-running work. Anthropic says the feature now works everywhere Claude does, including regular chat and Cowork, rather than treating those surfaces as separate contexts. That is particularly relevant for users who research in chat and then hand off execution to Cowork. (Anthropic)

Anthropic describes memory as a collection of topics, not an uneditable transcript. Claude can organize saved context into topic files, and users can ask what it remembers, correct it, remove individual topics, or clear the memory entirely. (Claude on X; Anthropic)

User controls

The practical controls are:

• Review: Ask Claude what it remembers or inspect the available memory topics.
• Correct: Tell Claude when a saved detail is wrong or out of date.
• Remove: Delete an individual topic without clearing everything else.
• Clear: Remove all saved memory when starting fresh.
• Disable: Turn memory off when a task should not use saved context.

Anthropic says sensitive topics are off by default. Users should still review memory before using Claude for work where stale context, confidential details, or an incorrect preference could change the result. (Anthropic)

Plan and workspace controls

Anthropic says memory is available by default for Free, Pro, and Max users, while Team and Enterprise workspaces can manage the feature through administrator controls. In managed workspaces, topic files can give users a clearer view of the information being used to personalize their conversations. (Anthropic)

That division matters for teams. A personal user can decide what context is useful, but an organization may want memory disabled or bounded by workspace policy. Administrators should review the control before treating cross-surface continuity as a default for sensitive projects. (Anthropic)

What users can do now

1. Open Claude's memory settings and check whether memory is enabled.
2. Review the topics Claude has saved and remove anything stale or unnecessary.
3. Start a small, low-risk workflow in chat, then continue it in Cowork.
4. Ask Claude to summarize the context it is using before a consequential task.
5. For Team or Enterprise work, confirm the administrator's policy before relying on memory.

The update is most useful for repeat workflows: recurring research, ongoing content projects, and tasks that move between conversation and execution. It does not remove the need to verify the context Claude uses, especially when a project has changed since the information was saved.

Sources

• Claude by Anthropic on X, “Memory works everywhere and you decide what’s in it,” August 25, 2026: https://x.com/claudeai/status/2092299704864284888
• Anthropic, “Claude’s memory works everywhere—and you decide what’s in it”: https://claude.com/blog/claudes-memory-works-everywhere-and-you-decide-whats-in-it