Claude gets Memory: Anthropic adds Context Recall and Incognito Chats 

Claude AI

Anthropic is rolling out “memory” to its Claude app, giving the AI the ability to retain context about users, projects, and workflows over time. The company says the feature will help teams cut down on repetitive explanations and keep complex projects moving without constant resets.

Memory is now available to Team and Enterprise plan users, with administrators able to disable it at any time. Incognito chats—conversations that don’t save to history or memory—are rolling out to all Claude users.

According to Anthropic, memory is designed for professional use cases. It can remember sales pipelines, product specifications, client details, and ongoing initiatives. Each project gets its own memory, keeping product launches, client accounts, and internal planning separate. This structure is meant to help teams juggle multiple priorities while containing sensitive information within the right context.

Users can view, edit, or delete what Claude remembers through a “memory summary” in settings. The AI updates its references based on user instructions, giving teams a measure of control over what information persists.

Incognito chat, meanwhile, offers a clean slate for sensitive or temporary work. Anthropic positions it as useful for brainstorming, confidential strategy, or simply starting fresh without the weight of prior context. Regular conversation history and memory remain unaffected.

The company stressed that memory comes with new safety considerations. To that end, it is limiting availability to workplace settings at launch and will expand more broadly after further testing.

How to use the Memory feature

Getting started is simple: enable memory in settings, allow Claude to build context from past chats, and try prompts like “what were we working on last week?” Anthropic also supports exporting or migrating memory data between tools.