Claude Cowork finally remembers what you told the app in chat
One of the most persistent pain points in the world of AI agents—the repetitive, tedious cycle of re-briefing your assistant on context it should already possess—is finally being addressed. Anthropic announced on Tuesday a significant architectural update that merges the memory systems used by its standard chat interface and its Claude Cowork agent.
This integration ensures that Claude maintains a consistent knowledge base across both environments. Whether you are brainstorming in a chat window or executing tasks in Cowork, the AI will now retain information learned in one session and apply it to the next, effectively transforming Claude from two distinct tools into a single, cohesive assistant.
Bridging the Gap Between Research and Execution
Previously, users often found themselves in a frustrating loop. You might spend hours developing a project strategy through deep, iterative conversations with Claude, only to find that when you transitioned to Claude Cowork to actually implement those ideas, the agent lacked the necessary background. You were forced to manually re-input project details, goals, and constraints.
With this update, that friction is removed. Anthropic highlights a common use case: drafting a conference agenda or a management update. Because the system now shares memory, Claude will automatically recall specific details—such as speaker names, venue locations, or headcount—without requiring the user to repeat themselves.
"The update makes Claude now feel like one continuous assistant, rather than two separate products under one roof."
Real-Time Memory Updates
Beyond simply merging the two platforms, Anthropic has overhauled how memory is captured. Claude will now update its memory store in real-time as you chat, rather than waiting for a conversation to conclude to generate a summary. This ensures that the information flow between chat and Cowork is instantaneous, allowing for a more fluid, responsive workflow.
Furthermore, users now have full transparency and control over their data. Anthropic is exposing the "memory bank," allowing you to:
- View exactly what Claude has retained about you or your projects.
- Edit specific details to ensure accuracy.
- Delete information that is no longer relevant or desired.
Privacy and Sensitive Data Handling
Anthropic has implemented clear guardrails regarding what the AI is permitted to store. By default, the system is designed to exclude highly sensitive personal data, including:
- Health and medical information.
- Religious, political, or philosophical beliefs.
- Race, ethnicity, or gender identity.
Users who prefer to have the AI remember these details can manually toggle the “include sensitive topics in memory” setting. Regardless of user preference, the system maintains strict prohibitions against saving high-risk data, such as Social Security numbers, government-issued IDs, or information related to criminal history.
This feature is rolling out immediately across Free, Pro, and Max plans on web, desktop, and mobile platforms. Users on iOS and Android should ensure they have updated to the latest version of the app to access these improvements.