Key facts
- Anthropic has merged the memory systems of Claude's chat and Cowork features.
- This integration allows Claude to retain information learned in one context across all interactions.
- Users can now access, modify, or remove data stored in Claude's memory.
- The update aims to provide a more continuous and seamless AI assistant experience.
- Claude will now add topics to memory during ongoing chats, rather than only at the end.
Anthropic has announced a significant update to its AI assistant, Claude, integrating its chat and Claude Cowork memory systems. This change aims to eliminate the frustration of repeatedly providing the AI with information it has already learned, creating a more cohesive user experience.
Previously, users might develop project ideas through conversations with Claude's chat function, only to find that this context was not carried over when using Claude Cowork for action-oriented tasks. The new unified memory system ensures that knowledge gained in one area is accessible in others, making Claude function more like a single, continuous assistant.
For example, when drafting a manager update or conference agenda, Claude Cowork will now automatically recall relevant details such as conference headcount, city, and speakers, based on prior chat conversations. This eliminates the need for users to re-input such information.
Furthermore, Claude will now add topics to its memory in real-time as conversations progress, rather than waiting for a chat to conclude. This allows for quicker updates and seamless sharing of information between chat and Cowork, even while a conversation is still active.