Anthropic just flipped the memory switch for Claude, finally bringing its AI chatbot up to speed with ChatGPT and Gemini. Starting today, paid subscribers can turn on Claude's memory feature, letting the bot remember past conversations without constantly being prompted - a capability rivals have had for over a year.
Anthropic is finally closing the memory gap. The company's Claude chatbot is getting a major upgrade that lets it remember past conversations automatically - no more asking it to please remember what you just discussed five minutes ago. Max subscribers can flip the switch today, while Pro users will see the rollout "over the coming days," according to Anthropic's announcement. The company stayed quiet on whether free users will ever get this feature.
The timing isn't coincidental. Claude has been playing catch-up in the memory game while OpenAI and Google have been letting their chatbots remember user preferences and conversation history since early 2024. It's the kind of sticky feature that keeps users coming back to the same platform instead of starting fresh with a competitor every time.
"Complete transparency" is how Anthropic describes its approach. Unlike the "vague summaries" you might get elsewhere, Claude will show users exactly what it remembers. You can toggle specific memories on and off through natural conversation - tell it to "forget an old job entirely" or focus on particular details. The company's also adding "distinct memory spaces" to prevent your work projects from bleeding into personal chats.
The competitive pressure is real. ChatGPT and Gemini have been using memory features as user retention tools, making it painful to switch platforms when you'd lose all that conversational context. According to reporting from The Verge, OpenAI rolled out ChatGPT memory in February 2024, while Google followed with Gemini memory features later that year.
Claude only gained basic conversation recall this August, but even then users had to explicitly ask it to remember things. That's a friction point Anthropic is clearly trying to eliminate. The company's also making it easier to switch by allowing memory imports from ChatGPT or Gemini through copy-and-paste, plus promising "no lock-in" with anytime memory exports.
