Claude's getting memory for Pro and Max subscribers, which means you can finally stop re-explaining your projects every time you open a new chat. The feature's live now for Max users and hitting Pro accounts over the next few days.
Key Points:
- Memory available now for Max users, rolling to Pro subscribers over coming days—eliminates repeating project details and preferences
- Each Claude Project gets its own isolated memory space, preventing work context from mixing with personal chats
- Import/export tools let users move memory data between AI assistants, reducing lock-in
Here's what's different: Anthropic built memory around Projects. Each one gets its own isolated memory space. Your product launch planning stays separate from client work. Confidential discussions don't leak into general operations. It's not one giant memory blob that mixes everything you've ever told the AI—it's compartmentalized by default.
You control what Claude remembers through a summary view in settings. Edit it directly, tell Claude to forget specific things, or use Incognito mode for conversations that shouldn't be saved at all. The company's also shipping import/export tools, so you can pull your memory data from ChatGPT or Gemini and move it into Claude—or take it with you if you leave.
That portability matters. ChatGPT's had memory since April 2024 (now on the free tier), and Gemini added it in February 2025. Memory creates stickiness—once an AI knows your preferences and context, switching tools means starting over. Anthropic's betting that making it easy to move that data around reduces the penalty for being late.
The project-based approach is the cleaner design. ChatGPT and Gemini run memory across everything by default, which works until you're mixing professional strategy discussions with weekend hobby chats. Context collapse is a real problem—one developer complained ChatGPT kept referencing his fondness for "dogs wearing pelican costumes" in serious work prompts because it remembered a joke from weeks earlier.
Memory's baseline now for AI assistants. Anthropic shipped it 18 months after ChatGPT, but with tighter boundaries and data portability. If you're on Claude Pro or Max, turn it on if you want to stop typing the same context repeatedly. If you're evaluating tools, the feature gap just narrowed—now it's about whose implementation makes more sense for how you actually work.