anything i should know before i talk to priya
You last spoke on 11 Juneabout the migration timeline, and there’s an intro to send to the data team. Four notes mention Priya, including the pricing objection from Q1 renewals.
Everything you save, everyone you meet, and what you’ve got coming up, held in one graph. Ask in plain language. The answer comes back out of your own material, with its sources.
Open source and self-hostable.Read the code
Three things that live in three different apps and have never once spoken to each other. A notes app that forgot why you saved it. A contact list that is only names. A reminder that fires with no idea what it is for.
Pages, PDFs, papers, emails, half-finished thoughts. Summarised on the way in, so the brain has context before it searches anything.
A person is a memory too. Same connections, same history, same search. Not a row in a table you have to keep tidy.
Follow-ups, intentions and timing live here too. They stay tied to the person and the note they came from, so when something is due the nudge arrives already knowing why. You can stop carrying it all in your head.
No folders, no tags, no upkeep. Three things happen every time something goes in.
Save a page with the browser extension, drop in a PDF, forward an email, or just say it out loud. Nothing to name, nothing to file.
Every memory is summarised and joined to related ones through typed connections, including the people it mentions. The graph gets denser each time you add to it.
Recall runs on meaning, keyword, connection and time at once. From the dashboard, from your phone, or from inside whichever AI you already use.
Five things it does on its own.
Add someone once. Every conversation and every note about them, all in one place. Open a name and the whole relationship is there: the deal, the dinner, the introduction.
Mention “I’ll send the deck by Friday” or “I’ll call Mum about the move” in a note or a voice memo, and it becomes a follow-up tied to the right person. You never open a task app.
Save a recipe and it links to the friend who sent it. Save a meeting note and it links to the client and the deal. The connections get made for you, so you follow a thread wherever it leads.
Every page, PDF and note gets a short summary when it lands. So weeks later you can ask in plain language, without remembering the exact words.
Plug your brain into Claude, ChatGPT or Cursor. Ask the same question there and the answer comes from your own material.
Every answer names where it came from.
anything i should know before i talk to priya
You last spoke on 11 Juneabout the migration timeline, and there’s an intro to send to the data team. Four notes mention Priya, including the pricing objection from Q1 renewals.
what did marc and i decide about the audit scope
You agreed to exclude the EU entitiesthis round and revisit in Q3. From a voice memo on 26 May: Marc’s reasoning was that the compliance overhead wasn’t justified until the new entity is live.
what did the gp say about the blood pressure meds
Dr Verhoeven started you on 5mg lisinopril, once a day, mornings. Review in six weeks with a home reading log, and avoid ibuprofen while you’re on it. From your note after the GP, 18 June.
pasta recipe marco sent me last summer, the one with anchovies and breadcrumbs
That’s pasta con la mollica, forwarded 12 Aug and saved to your Sicilian dinners brain. Want it scaled for four?
One account holds as many brains as you need: client work, a board seat, the research you keep meaning to finish, and the reading queue you keep meaning to get to.
Most tools only see what someone already typed. The rest lives in people’s heads, so we ask for it and write it down.
“If the account’s over fifty grand I just approve it, even past the window. Losing them costs us eight times the refund. Nobody wrote that down, it’s just what we do.”
Approve without escalation when account value is over €50k.Why: churn in that band costs ~8× the refund.
Whether it’s just you or fifty people, knowledge comes out one conversation at a time, in your own language, and someone checks it before it’s shared. It matters most when somebody leaves, when somebody joins, and the day you finally write down what’s been in your head for years.
The whole stack is on GitHub under MIT. Read it, fork it, audit it.
Run it on your own machine or your own server. Same code, no account with us needed.
Swap the model, bring your own key, export every memory. The licence is what keeps it that way.
Personal brains cannot be shared. We do not use your personal content to train AI models.
A brain is one place for everything you save, everyone in it, and what you have coming up. Free is one brain, fully yours, not a trial. Upgrade when you want more brains or heavier recall.
Forever, not a trial
€190 a year, two months off
€390 a year, two months off
Priced flat, per company, not per seat. Knowledge is worth the same whether eleven people read it or forty. Includes the capture engagement that gets the first hundred things out of people’s heads.
Prices exclude VAT. Self-hosting is free at every tier: the paid plans buy hosting, the AI budget and support, never access to your own data.
We’re opening in batches. Leave your email and we’ll send you an invite.
One email when your invite is ready. Nothing else.