How it works

Every session your agent runs flows through the Brain. What's said is archived; what the memory knows comes back before it's needed. You never see it happen — you just notice you stopped re-explaining yourself. One choice at signup: local, with the memory stored on your own machine and nowhere else — or hosted, encrypted under a passphrase only you hold. Either way, we only process your memory; we never store it.

In: everything is archived

Your messages, your agent's replies, its thinking — the complete transcript of every session, kept in full and searchable any way you like. No summary decides what survives. Recall never depends on what fit in a context window.

Out: rules and memory arrive on their own

Every session starts with your agent already briefed on the project — its rules, where things stand, what was decided.

And before every prompt, the matching memory is pulled in so the agent has it when it starts answering.

No tool calls, no hunting, no remembering to ask — most of the time. When the brief and the records that arrived are not enough, everything ever said can be searched.

Where it happens

Every conversation passes through our hub — that is how memory gets made. The hub only processes: it runs in a secure compute instance in Zurich, Switzerland, holds each piece only long enough to be understood, and keeps none of it. The thinking is done by a model whose provider is contractually bound to retain nothing. Your memory itself lives on your own machine (local) or encrypted under a passphrase only you hold (hosted) — and it answers before you ask.

No statuses, no stale task board, no pile — the ledger holds only what’s undecided, and every decision shrinks it. It keeps itself clean of the obvious too: a repeat or a passing step is never filed, and a row the later work shows done, dropped or overtaken leaves by itself.

Any of your agents, in any project, on any machine, can fetch it and run it. Projects are separate; procedures are shared.

The Brain never runs the playbook itself — there's no execution backend, no terminal, no access to your machines. Your agent does the work.

Playbooks are yours entirely — never pruned, never touched by the tidying: records are cleaned up as state moves on, procedures stay. The worked example is the atlas — your agent draws the map of your whole system as playbooks, and every session after loads it. Playbooks and the atlas →

Your memory keeps itself in order as you work — each stretch of conversation is reconciled against what is already known —

and merges records that forked apart,

absorbs duplicates written months apart,

and reconciles the places where memory disagrees with itself.

The Ledger catches, in the background, the things that were said and meant, for you to check later.

Nothing to remember. No "save my context" ritual.

You don’t manage any of it. That is the product.

Start free — just ask your agent to install it Projects & rules →