Hi PH! We built hippocampOS because we got tired of re-explaining ourselves to our AI every single morning. Every session starts blank...your role, your projects, the decision you made on Tuesday. You type the same paragraphs again, or re-upload the same files, and then you finally ask your actual question. If you use Claude or ChatGPT seriously, this is a real daily tax and it never stops. hippocampOS builds a private memory graph from where your work actually lives: your email, meetings, notes and files. Then it splits that into separate brains, one for each role you play: Work, A specific client, Research, Personal. You invoke one by name inside Claude or ChatGPT and your context arrives in under a second (and with a fraction of your AI's usual token usage in 'Projects'.) The question everyone asks first: why not use Claude's built-in memory? Because it remembers your chats, not your work. It only knows what you happened to tell it, it's one It only knows what you happened to tell it, it's one blurry profile rather than sealed brains, and it lives inside one vendor. Keep it on. hippocampOS covers what it can't see, and it comes with you when you switch models. Two things we care about: 1. Your brains are sealed. Client A's context never surfaces in client B's chat. For anyone juggling multiple clients, this is why we built this 2. You don't have to hand us your email or meeting notes to start. Paste our Brain Seed prompt into Claude, let it summarize what it already knows about your work, upload the result. Working brain in five minutes, nothing sensitive shared. Connect real sources when you're ready. About us: there are two of us, both with day jobs we like, working with AI we constantly have to introduce ourselves to every day! hippocampOS is encrypted at rest and in transit, never trained on, delete means delete, and you can export everything as markdown whenever you want. From $5/mo. Pay at signup, full refund within 7 days--just send us one email, no questions asked. We'll both be here all day! Ask us anything...we can share results from memory quality benchmarks, use cases we love, or anything else. If you try it and it doesn't work for you, tell us that too. That's tbh more useful to us right now than an upvote! hippocampos.io (thank you!)

hippocampOS

Give your AI a memory of your work

Give your AI a memory of your work