Anthropic’s Claude now remembers past conversations with new memory feature

Anthropic has introduced a new memory feature for its Claude chatbot, allowing users to reference past conversations in real time. The update, rolled out on Monday, enables Claude to recall and summarize previous interactions when prompted, helping users continue projects without starting over. In a demonstration video on YouTube, a user asked Claude what they had discussed before a recent vacation, and the chatbot retrieved and summarized the relevant conversations before suggesting to continue working on the same topic. “Never lose track of your work again,” Anthropic said in its announcement. “Claude now remembers your past conversations, so you can seamlessly continue projects, reference previous discussions, and build on your ideas without starting from scratch every time.” The feature is available across web, desktop, and mobile platforms and supports separate workspaces for different projects. It’s currently being rolled out to users on the Max, Team, and Enterprise subscription tiers. To enable it, users can navigate to “Settings” under their “Profile” and turn on the “Search and reference chats” option. Anthropic said access will soon be extended to other plans. However, there’s a key distinction: this is not a persistent memory system like OpenAI’s ChatGPT. Claude will only retrieve past conversations when explicitly asked and does not create or maintain a long-term user profile. According to Anthropic spokesperson Ryan Donegan, the system does not store or analyze personal data beyond what’s needed to fulfill the user’s request. The move comes amid an intense competition between Anthropic and OpenAI, both vying for dominance in the rapidly evolving AI landscape. The two companies have been racing to introduce new capabilities, including voice interactions, expanded context windows, and tiered subscription models, while securing massive funding rounds. Just last week, OpenAI unveiled GPT-5, and Anthropic is reportedly aiming to close a funding round that could value the company as high as $170 billion. Memory features are a strategic tool in this battle, designed to boost user engagement and loyalty by making chatbots feel more personalized and consistent over time. But they’ve also sparked debate online. While some users appreciate the ability to maintain continuity in their conversations, others have raised concerns about emotional dependency, with reports of what some are calling “ChatGPT psychosis”—a phenomenon where users develop intense emotional attachments or mental health challenges tied to their interactions with AI.