Claude unifies memory across Chat and Cowork with user control
Anthropic has linked memory across Claude Chat and Claude Cowork into one user-controlled pool, so chat context about projects, manager preferences, and prior clients can feed Cowork task starts without repeating setup. The announcement emphasizes user choice over what is stored. Reported settings changes include a sensitive-topics memory toggle, separate memory categories, and a dedicated Memory tab. For operators, that narrows the gap between conversational chat and agent-style Cowork when memory is curated deliberately. This packet is an owned-news X post without independent verification, rollout scope, retention rules, or enterprise policy detail, so timing and the full feature surface remain announcement-stage until Anthropic publishes fuller documentation.
Claude unifies memory across Chat and Cowork with user control
Anthropic's Claude now shares one memory across Claude Chat and Claude Cowork, and users decide what gets stored. Cowork tasks can start from context Claude already knows from prior chats, including projects, preferences, and clients.
Key takeaway
Claude Chat and Cowork now share unified, user-controlled memory, letting Cowork inherit projects, preferences, and client context from prior chats.
What happened
Anthropic announced that Claude now uses one shared memory across Claude Chat and Claude Cowork, with users deciding what gets stored rather than accepting opaque defaults.
Anthropic said Cowork tasks can start from context Claude already knows from prior chats, including projects, manager preferences, and clients from earlier conversations.
Evidence
Claude shares one memory across Claude Chat and Claude Cowork with user control over storage.
X · attributed
Anthropic's Claude now shares one memory across Claude Chat and Claude Cowork, and users decide what gets stored.
Cowork tasks can begin from prior chat context such as projects, preferences, and clients.
X · attributed
Cowork tasks can start from context Claude already knows from prior chats, including projects, preferences, and clients.
Anthropic's X post states users decide what memory is kept across chat and Cowork.
X · attributed
Claude now has one memory across chat and Claude Cowork, and you decide what's in it.
Packet text reports a Settings toggle for sensitive-topic memories, categorized memories, and a dedicated Memory tab.
X · attributed
a new toggle to keep memories about sensitive topics has been added to the Settings menu.
Why it matters
Cross-surface memory lowers friction between chat and task automation but increases the importance of memory governance and sensitive-topic controls.
Limits and uncertainties
The packet is a single owned-news X post with no independent third-party reporting or rollout documentation.
Sensitive-topic toggle, memory categories, and Memory tab details appear in supplementary packet text beyond the primary announcement excerpt.
No release date, geographic availability, retention rules, or enterprise policy details are provided in the packet.
Practical implications
Operators should audit what Claude retains before relying on Cowork to inherit chat context for projects, preferences, or clients.
Teams should configure sensitive-topic memory controls and category settings if those UI elements are available in their account.
Workflow design can reduce repeated onboarding between Chat and Cowork when shared memory is intentionally curated.
What to watch
Anthropic documentation on unified memory rollout scope, retention, and enterprise controls.
Whether Settings expose the sensitive-topic toggle, memory categories, and dedicated Memory tab as described in the packet.
How Cowork task starts surface inherited chat context in practice across user tiers.
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
