Grounded
Watch scattered apps compile into one connected memory.


Your life is scattered across twenty apps.
Ostler reads them all and works out who is who – on your Mac.
Into one connected page for every person – every fact cited to its source.
Your life is scattered across twenty apps.
Ostler reads them all and works out who is who – on your Mac.
Into one connected page for every person – every fact cited to its source.
No single-source app can answer “when did I last see Rae, and what did we talk about?” A memory built only from your calendar knows when you met someone. A memory built from your meetings, messages, notes and reading together knows who they are to you.
Ostler reads what is already on your machine – email, messages, WhatsApp, iMessage, calendar, browsing history, notes, AI conversations – and fuses it into one graph. This, not “on-device” in the abstract, is the hard, differentiated thing. It is also the thing a chat-only assistant structurally cannot do: it only ever had the chat.
Imagine a librarian who works only for you. She reads everything you already have – a decade of email, the group chats, the meeting notes, the articles you saved and forgot – and writes the encyclopaedia of your life: a page for every person you know, cross-linked to every meeting you have had with them; a page for every topic you care about; timelines, heatmaps, the lot. She updates it while you sleep.
That encyclopaedia is your personal wiki – tens of thousands of interlinked pages, compiled automatically. It is the visible tip of the real asset underneath: a personal knowledge graph fused from every source. The wiki is how you browse it; the assistant is how you question it.
A chatbot knows what you typed today. Ostler knows what you lived.
Journaling apps start empty and die of the blank page. Ostler starts full: it backfills from the archive you already have, so within days of install you have a wiki with thousands of pages about a life you have already lived. Nobody else’s first run can do this, because everyone else asks you to start generating data. Ostler asks only for permission to read what exists.
Reading a decade takes time – you get your first surprising page in minutes, and the deep past fills in over a few days, with the progress shown plainly as it goes.
The second brain people spend years building by hand – except it fills itself in.
Want the exact boundary between what Ostler imports, what it infers, and what it never stores? That is laid out on what Ostler knows about you. And the reason a memory this deep is a comfort rather than a liability is where it lives.