Anthropic just made its memory feature meaningfully more useful, and the key word there is ‘meaningfully.’ Not a rebrand, not a benchmark. An actual behavioral change that affects how much you have to repeat yourself when using Claude day to day.
As reported by Engadget, Anthropic has updated Claude so that memories now carry across both the standard chat interface and Claude Cowork, its computer-use agent that automates repetitive tasks. Before this update, those two surfaces operated in separate memory silos. If you told Claude something in a chat session, Cowork had no access to it. That friction was a real problem for anyone trying to build a consistent working relationship with the model.
Cowork launched earlier this year and lets users automate routine tasks that run in the cloud. The pitch is that it handles the repetitive stuff while you focus on higher-level work. But an agent that can’t carry over context from your regular conversations is a limited agent. This fix addresses that directly. “When Cowork runs a task in the cloud, what Claude remembers from your chats is there, and vice versa,” Anthropic explained. The timing also connects to a recent update for the Claude for Chrome extension, which made Cowork accessible through the browser plugin and folded Chrome side panel conversations into your usage history.
The update also adds a new “include sensitive topics in memory” toggle. By default, Claude skips storing anything related to health, race, ethnicity, religion, politics, or gender identity. Flip the toggle, and Claude will start saving those topics, with a notification each time it does. You can turn it off at any time and edit or delete anything it has stored. Notably, some data stays off-limits regardless of your settings:
- Social security numbers
- Criminal history
- Immigration status
Anthropic’s reasoning for adding the toggle is straightforward: what counts as sensitive varies by person. That’s a fair point, and it’s a more honest approach than pretending one default works for everyone. The company is also not applying these settings retroactively, which matters for users who care about what’s already in the system.
This fits into a broader pattern across the AI assistant space. OpenAI has offered memory in ChatGPT for some time, and Google is building persistent context into Gemini. The competition on memory is less about whether it exists and more about how well it works across an entire product ecosystem. Anthropic is now making a clearer case that Claude’s memory isn’t just a chatbot feature but a foundation for the whole suite.
Memory is on by default for Free, Pro, and Max accounts. Mobile users need to update to the latest Claude app on Android or iOS to get the new options.
