Anthropic announced a major update to its Claude AI assistant today, introducing persistent memory across conversations and expanded tool use capabilities. The new features allow Claude to remember user preferences, past interactions, and project context across multiple sessions, while also enabling the model to interact with external software tools and APIs on behalf of users.
The persistent memory feature stores user-approved information in an encrypted personal context that Claude can reference in future conversations. Users maintain full control over what is remembered and can view, edit, or delete stored memories at any time. Anthropic emphasized that memory data is encrypted and never used for model training.
The tool use capabilities allow Claude to browse the web, execute code, manage files, and interact with third-party applications through a secure sandboxed environment. Anthropic CEO Dario Amodei said the features represent a shift from conversational AI to what the company calls collaborative AI, where the assistant can take meaningful actions in the real world while maintaining the safety guardrails that Anthropic is known for. The features are rolling out to Claude Pro subscribers this week.
