Anthropic unveiled a major update to its Claude AI suite on Tuesday, merging the memory architecture that powers both the chat‑based Claude and the productivity‑focused Claude Cowork. The integration lets the assistant retain details from one interaction and apply them in another, removing the repetitive step of re‑explaining project context when moving from brainstorming to execution.
Previously, users who built a project outline in Claude’s chat window discovered that the same information was not automatically available to Cowork. They had to repeat headcounts, venue details, or speaker lists when transitioning to task‑oriented workflows. The new shared memory eliminates that friction, making the two products feel like a single, continuous assistant.
Beyond seamless context switching, Anthropic is giving users direct access to what Claude has stored. A memory dashboard lets individuals read, edit or delete any topic the AI has retained. By default, the system excludes personal or sensitive categories such as health information, race, religion, gender identity and political views. Users who wish to include those topics can toggle an "include sensitive topics in memory" option, and the platform will flag each instance when it saves such data.
Anthropic emphasizes that certain categories will never be stored, regardless of user preferences. Government‑issued IDs, Social Security numbers, criminal histories, immigration status and other data that breach its acceptable‑use policy are permanently blocked from memory.
The update is enabled by default on all Anthropic plans—Free, Pro and Max—and is available across web, desktop and mobile interfaces. iOS and Android users must install the latest app version to activate the feature. According to the company, the memory now updates in real time as conversations progress, rather than waiting for a session to end before summarizing key points.
Industry observers note that the move aligns Claude more closely with competitors that already offer persistent context, potentially raising the bar for user expectations of AI assistants. Anthropic’s approach to privacy—allowing granular control while defaulting to exclusion of sensitive data—offers a middle ground between functionality and user protection.
Early adopters have reported smoother workflows. One user described drafting an agenda for an upcoming conference: Claude remembered the event’s city, speaker lineup and attendee count without needing to re‑enter the details in Cowork. Another developer highlighted that the shared memory cut down on repetitive prompts, freeing time to focus on higher‑level tasks.
Anthropic says the memory feature will continue to evolve, with future updates aimed at refining relevance algorithms and expanding user‑controlled filters. For now, the integration marks a notable step toward making AI assistants feel truly collaborative rather than compartmentalized tools.
Dieser Artikel wurde mit Unterstützung von KI verfasst.
News Factory APP - agentische News für besseres SEO & AEO.