Claude unifies memory across Chat and Cowork with user control
Anthropic said Claude now uses one shared memory layer across Chat and Cowork, letting users choose what is stored rather than relying on separate histories per surface. When a Cowork task starts, Anthropic says it can draw on context already gathered from prior chats, such as ongoing projects, manager preferences, or past clients, so hand-offs begin from known facts instead of a blank slate. Reported UI changes include a dedicated Memory tab, category splits for topics including sensitive ones, and a new Settings toggle governing whether memories about sensitive subjects are retained. For builders and operators, the shift is less about a single feature flag and more about how cross-surface persistence and explicit user controls reshape assistant continuity, with rollout details and retention policies still to be confirmed from Anthropic's announcement alone.
Claude unifies memory across Chat and Cowork with user control
Anthropic says Claude now shares one memory across chat and Claude Cowork, and users decide what is stored. Cowork tasks can start from context Claude already knows from prior chats, including projects, preferences, and past clients.
Key takeaway
Claude Chat and Cowork now share one user-controlled memory, so Cowork tasks can start from prior chat context instead of isolated sessions.
What happened
Anthropic says Claude now shares one memory across chat and Claude Cowork, and users decide what is stored rather than accepting automatic retention by default.
Anthropic says Cowork tasks can start from context Claude already knows from prior chats, including projects, preferences, and past clients, and reporting also notes a dedicated Memory tab, category splits for sensitive topics, and a Settings toggle for keeping memories about sensitive subjects.
Evidence
Claude shares one memory across chat and Claude Cowork with user control over what is stored.
X · attributed
Anthropic says Claude now shares one memory across chat and Claude Cowork, and users decide what is stored.
Cowork tasks can start from prior chat context such as projects, preferences, and past clients.
X · attributed
Cowork tasks can start from context Claude already knows from prior chats, including projects, preferences, and past clients.
Claude added a Settings toggle for keeping memories about sensitive topics.
X · attributed
a new toggle to keep memories about sensitive topics has been added to the Settings menu.
Memory topics are split into separate categories and Memory has a dedicated tab in Claude.
X · attributed
Memoriesed topics are now split into separate categories. Memory now has a dedicated tab in Claude
Why it matters
Unified, user-governed memory across Chat and Cowork changes how persistent context flows between conversational and task-oriented workflows inside Anthropic's product surface.
Limits and uncertainties
The packet is based on Anthropic's X announcement and related snippets, without independent technical documentation on storage scope, retention defaults, or rollout timing.
Details such as how sensitive-topic categories are defined and enforced are not fully specified beyond the reported Settings toggle and category split.
Practical implications
Teams using both Claude Chat and Cowork should review what memories are stored before handing Cowork tasks that rely on prior chat context.
Operators should locate the new Memory tab and sensitive-topic toggle in Settings to align retention with policy before sharing client or manager-specific details in chat.
What to watch
Whether Anthropic publishes fuller documentation on unified memory retention, deletion, and sensitive-topic controls beyond the X announcement.
How Cowork task quality changes when it starts from shared chat memory for projects, preferences, and past clients.
Original reporting: Claude now has one memory across chat and Claude Cowork, and you decide what's in it.
Hand Cowork a task and it starts from what Claude already knows from your chats: the project
