Artificial intelligence company Anthropic has officially launched a new "memory" feature for its flagship AI model Claude's Pro and Max versions. This feature aims to significantly enhance the consistency and personalized experience of user conversations.
Core Features and Privacy Design
Through Claude's memory feature, the system will be able to remember details such as project content, user preferences, and work processes, ensuring continuity in subsequent conversations. Anthropic has particularly focused on privacy and isolation in the design:
Project-specific memory: Each project has an independent memory function, meaning that confidential topics or content from different fields can remain independent and not interfere with each other.
User control: Users can view and edit the content of Claude's memory at any time.
"Invisible Chat" mode: The newly added "Invisible Chat" option allows users to have conversations that do not save data or appear in chat records.
This memory feature is optional, and users can choose to enable it in the settings.
Security and Testing First
Before the release of the feature, Anthropic stated in a blog post that they conducted rigorous testing on sensitive areas of the memory storage function to ensure that it does not reinforce harmful conversation patterns or bypass any security features, and made necessary adjustments based on the test results.
Notably, this memory feature was already available to team and enterprise version users
