Anthropic puts persistent memory into Claude Cowork

Anthropic has added persistent memory to Claude Cowork, enabling the AI assistant to retain context from previous conversations and pick up where users left off without re-prompting. The memory feature is on by default, but saving sensitive topics is off by default, and Claude will not store sensitive PII such as Social Security numbers, government IDs, criminal history, or immigration status even if the sensitive topics feature is enabled. Users can manage memory via Settings > Memory, and changes to memory files carry over to all future conversations.

Anthropic puts persistent memory into Claude Cowork One of the big complaints engineers working with AI have is that it has no memory, so users basically have to start from scratch, re-prompting the same data and conversations it had shared with the model the day before. It was Groundhog Day for engineering. Today, Anthropic solves that issue by announcing the same memory it used in chat is now built into Claude Cowork. The result allows users to pick up where they had left off, and the context provided over months of conversations is saved. Ih a blog announcing the news https://claude.com/blog/claudes-memory-works-everywhere-and-you-decide-whats-in-it , Anthropic wrote: “Ask Cowork to draft an update for your manager, and it already knows who that is and how she likes updates written. Brainstorm the agenda in chat for a conference you’re organizing; when Cowork builds the budget and logistics doc, it knows the headcount, the city, and the speakers. Explain once in chat how your team defines its metrics, and every quarterly business review deck Cowork builds after that uses them, with no rebriefing.” The conversations Claude remembers are stored in files that can be read, edited or deleted under the Topics in Memory settings, the company said. Also, if the user needs to make a change or correction in one file, that change is carried over into every conversation from then on. Anthropic said the memory, though, will not store topics related to such things as health, ethnicity, race, gender identity and more, though some might consider these sensitive areas useful, and they can turn on an “include sensitive topics in memory” feature. But to protect safety and privacy, Claude will not stores sensitive PII like social security numbers, IDs, criminal history or immigration status even if the sensitive topics feature is turned on. Whenever Claude saves a sensitive topic to memory, a notice is sent to the user, and those topics are saved going forward and are not retroactive to earlier conversations. The new memory feature is on by default, but saving sensitive topics is OFF by default. Visit Settings Memory https://claude.ai/redirect/claudedotcom.v1.423207fc-bd03-4b0f-b66f-d84351c639ef/new settings/customize-memory to turn memory on and control what Claude remembers. SD Times Q&A Does Claude remember conversations across sessions? Yes. Anthropic has added persistent memory to Claude Cowork, allowing it to retain context from previous conversations. Users can pick up where they left off without re-prompting, as the model stores context accumulated over months of interactions. What sensitive data does Claude Cowork refuse to store in memory? Even if the ‘include sensitive topics in memory’ feature is enabled, Claude will not store sensitive PII such as Social Security numbers, government IDs, criminal history, or immigration status. Topics like health, ethnicity, race, and gender identity are excluded from memory by default. How does Claude Cowork notify users when sensitive topics are saved to memory? Whenever Claude saves a sensitive topic to memory, it sends a notice to the user. Sensitive topic memories are only saved going forward and are not applied retroactively to earlier conversations.