Anthropic has launched a new memory feature for its chatbot Claude, which allows it to recall and summarize previous conversations.
In the demo video, a user asked what they had talked about with Claude before their vacation, and the bot then reviewed their chat history, summarized it, and suggested they continue working on the same project. According to Anthropic, this allows for continuous project management, reference past discussions, and development of ideas without having to start from scratch each time.
The feature is already available on the web, desktop, and mobile. It also allows you to save separate workspaces and projects. It is currently rolling out to Claude Max, Team, and Enterprise plans.
To activate the new feature, you need to enable the "Search and link to chats" option in your account settings. Access for other plans will be available later.
However, Anthropic emphasizes that this is not a persistent memory like OpenAI's ChatGPT. Claude does not build a user profile and only accesses past chats upon direct request.
The launch of this feature is another step in the competitive struggle between Anthropic and OpenAI, which are actively introducing new features such as voice modes, expanded context windows, and new tariff plans, while simultaneously attracting investment.
By the way, last week OpenAI released GPT-5, and Anthropic is preparing for a new round of funding, which could raise its valuation to $170 billion.