Picture a freelance graphic designer juggling five clients, each with a different brand voice and color palette. (She’s an illustration, not a real person.) Before this update, she could explain a client’s preferences in one Claude tool and find the other had no idea. On August 25, 2026, Anthropic announced a fix: the Claude memory feature now adds topics as you chat, and it’s shared between the regular Claude chat and Claude Cowork.
Why It Matters to You
If you keep explaining the same background to an AI assistant, shared memory means you may have to do it less. Freelancers, students and small business owners are who Anthropic’s examples point to. It’s on by default for Free, Pro and Max plans, and work accounts differ because admins decide whether memory is available. Anthropic hasn’t published any measure of time saved, so we don’t claim one.
What changed and what Claude keeps
Chat and Cowork used to keep separate memories. TechCrunch reports that someone who developed a project idea in chat had to explain it all again to Cowork. Anthropic merged the two systems and changed the timing too. Claude “now adds topics to memory as you chat, instead of summarizing conversations after they end,” and TechCrunch adds that memories sync faster, even while a chat is still open. Anthropic’s own example is a manager update or conference agenda, where Cowork already knows details such as headcount, city, and speakers from earlier chats.
You can read, edit, or delete what’s stored. Memories appear as short files under Topics in Memory settings, edits apply to future conversations, and you can pause memory or reset it. Sensitive topics are off by default, among them health, race, ethnicity, religious beliefs, politics, and gender identity. You can switch on “include sensitive topics in memory,” and Claude then shows a notice each time it saves one. That doesn’t save earlier conversations after the fact. Some items are never stored even with the setting on: sensitive ID numbers, criminal history, immigration status, and anything that breaks Anthropic’s acceptable use policy.
Plans, cost and the fine print
Anthropic says memory is on by default for Free, Pro, and Max plans on web, desktop, and mobile (mobile users need the latest app version). On Team and Enterprise plans, admins control availability and individuals start with memory off until they turn it on. The announcement lists no extra charge. As for limits: this is Anthropic’s own description of its product, and we found no independent testing of how reliably memory works. Memory only helps if what Claude stored is right, so check your Topics files now and then and fix wrong entries. Everything outside the sensitive categories is kept until you edit or delete it.
Do I have to pay to use Claude’s memory? Anthropic says memory is on by default for Free, Pro, and Max plans, and its announcement lists no extra fee.
Will Claude remember my health or religion? Not by default. Sensitive topics are off unless you turn them on, and Claude then shows a notice each time it saves one.
Sources
- Anthropic: Claude’s memory works everywhere, and you decide what’s in it (Aug 25, 2026)
- TechCrunch: Claude Cowork finally remembers what you told the app in chat (Aug 25, 2026)