Claude Cowork’s AI assistant finally remembers your past chats across the app

Anthropic has upgraded Claude’s memory system, merging the memory shared by its chat and Cowork experiences so what you tell the AI in one space sticks across others. This change means no more repeating over and over when shifting from brainstorming with Chat to executing in Cowork.

The system now captures topics during the conversation rather than waiting for the session to end, so chat and Cowork are aligned live. As you move from gathering ideas to taking action, Claude carries forward things like project background, city names, schedule details — whatever’s come up before — automatically.

Privacy & sensitive info controls

By default, Claude avoids storing highly private data like race, religious belief, or medical history unless you explicitly tell it to. If you opt in to “include sensitive topics in memory,” it will include such information. And it warns you when it’s doing so. Certain categories — government IDs, social security numbers, criminal or immigration status — are never saved, under any circumstances.

Availability & user experience impact

The memory feature is now live for Free, Pro, and Max users across web, desktop, and mobile platforms. People using iOS or Android will need to update the app to get the new behavior. With this change, Anthropic aims to make transitions between ideation and execution more seamless.

This update addresses feedback: earlier, Claude’s chat and Cowork felt like separate tools. With shared memory, your AI assistant behaves more like a unified helper, one that builds on what you’ve already told it.

Analytical Take: This move puts Claude in sharper competition with other AI agents that already carry forward user context — a core part of delivering personalized, efficient digital assistants. It raises real stakes around what information is stored and how, especially as AI becomes more embedded in workflows. Going forward, users should watch how Anthropic refines memory controls and transparency — because longer memory means greater benefits, but also higher responsibility.