Anthropic Merges Claude Chat and Cowork Memory

Update (9 Nov 2026): Claude Fable 5.2 and Mythos 5.2 are out - prices, access and benchmarks are in my Claude Fable 5.2 guide.

👁 118 reads

Anthropic announced on Tuesday that it is merging the separate memory systems used by Claude’s chat interface and its Claude Cowork agent, so that context learned in one area is automatically available in the other - no rebriefing required.

Before the change, a user could spend hours in chat building up shared understanding of a project - headcounts, locations, contacts - only to find that Cowork, the action-taking side of the product, had no knowledge of any of it. The update removes that wall. Anthropic gives the example of planning a conference: Cowork would already know the city, speakers, and headcount from prior chat, so drafting an agenda or manager update requires no re-entry of those details.

Alongside the unified memory, Anthropic is changing when memories are written. Previously, Claude summarised a conversation after it ended; now it adds topics to memory continuously as a chat progresses. That means Cowork picks up new context even while a conversation is still open, tightening the loop between research and action.

The update also gives users direct visibility into what Claude has retained. Memory contents can be read, edited, or deleted at any time. By default, Claude will not store sensitive categories - health data, race, ethnicity, religious beliefs, political views, or gender identity - though users can enable those by toggling on “include sensitive topics in memory.” Certain data, including government-issued ID numbers, Social Security numbers, criminal history, and immigration status, will never be stored regardless of settings. The app will notify users whenever a sensitive topic is saved.

Memory is enabled by default across Free, Pro, and Max plans on web, desktop, and mobile. iOS and Android users need to update to the latest app version to receive the feature.

Practitioner takeaway: Anyone who uses Claude for ongoing projects - content pipelines, client work, event planning - can now treat chat as a genuine briefing layer. Context built up over days of conversation will carry directly into agentic tasks without manual handoff. The ability to audit, edit, and delete memory also gives developers and business users a practical control point before sensitive or outdated information influences agent behaviour.

// read moreClaude Duet: Anthropic’s Proactive Agent ExplainedAI ToolsIs Claude Down? Live Anthropic Status & How to CheckAI ToolsAnthropic Revenue Hits $65bn as Fable Model Fails to Gain TractionMagazine

Kinetyra