- Claude’s new incognito mode is free for all users
- Conversations with the AI will stay private and not be in your history or memory
- The new features arrive with Claude’s newly upgraded memory system for subscribers to Claude Max, Team, and Enterprise
If you enjoy using Anthropic’s Claude AI chatbot but don’t really like the idea of your conversations lingering forever in the cloud, you’re in luck. Claude can now go Incognito, meaning any interaction will be private and unsaved. You won’t see it in your history or when you open the app.
In an industry where AI privacy often comes with a monthly price tag, Anthropic’s decision speaks to how explosively popular Incognito mode is among those leery of how much personal information digital tools absorb to be used at the whim of massive tech companies.
Claude now offers this kind of ephemeral, memory-free mode to every user on every subscription tier, including the free level. Just click the little ghost icon when starting a new chat, and it’s activated. The black border and label confirm your chat is incognito. When you close the window, it’s gone. No history. No memory. No trace aside from a temporary 30-day retention period for safety.
As with web browsers, Incognito mode is great if you want access to a digital toolkit without everything you look at being a potential news story. Maybe you are embarrassed by your personal, speculative, or just plain weird question. Now you can ask about it without the fear that Claude’s going to bring it up later or incorporate it into a future response. It’s not just about hiding embarrassing questions. It’s about giving users a mental sandbox: a space to think out loud, test ideas, or learn something new without it becoming part of the chatbot’s long-term memory.
That long-term memory has just started rolling out now for Claude. Unlike Incognito mode, though, the memory features are only for Team and Enterprise subscribers at the moment.
Opting into the memory features allows Claude to recall context from conversations, remember previous projects in Projects Mode, store notes about your work preferences, and even help you pick up where you left off. Each project’s memory is isolated, which means your work chats won’t bleed into your personal writing.
Claude remembers
But here’s where it gets interesting: incognito mode and memory don’t compete. They complement each other. Use incognito when you want a clean slate, free from influence or history. Use memory when you want Claude to be a continuity machine, helping you carry long-term threads across chats and tasks.
And if you’re the kind of person who changes your mind a lot about what you want remembered, Claude’s approach is refreshingly respectful. Nothing gets saved unless you opt in. And if you don’t want memory at all? You don’t have to use it.
It also sets Claude apart from some of its biggest rivals. While OpenAI’s ChatGPT and Google Gemini both offer their own versions of memory and private chats, they don’t make those distinctions quite as clear or customizable. Claude’s implementation feels unusually transparent thanks to its prominent labels and icons
Not having the memory feature in place is both appealing and seems to negate some of the possibilities of an AI chatbot. They’re their own bubble and =can’t be converted into regular ones after the fact, so if you forget to copy something important before closing the window, it’s gone. You also can’t use incognito mode inside Claude’s “Projects” feature.
Still, the broader implication that people want to at least have the option for privacy in their AI chatbot conversations is obvious. Incognito mode lowers the barrier to entry for people who are curious about AI but wary of leaving a data trail. And, oddly enough, an AI that can also forget things or at least imitate the experience seems a lot more human than one with total recall.
Add Comment