What changes when it remembers
Persistent memory isn't something you notice on day one. It's what makes day sixty different from day one.
Stop re-explaining yourself
Every session opens with who you are, what you're working toward, how you like to work and what changed since last time, before the first word of an answer.
The alternative isn't a longer prompt. It's the model quietly guessing at things you told it last week.
> start() today Sunday · first session of the day preferences how you like to work, in full protocols the rules you keep for the assistant threads 16 active · 0 dormant last session what was decided, and why changed 22 notes since you were here // the first answer already knows who you are // and what happened yesterday.
Every claim traces to a source
Creating a domain creates its Sources note in the same call. Each claim traces to a listed source, marked found or used, with a revision table recording when each was last checked.
BrainLLM ├── Sources every source, found or used, dated ├── Current State the latest measured values ├── Purpose and Design ├── Tool Surface ├── Runtime Architecture ├── Session Protocol └── ... each partOf the domain
Several agents, one memory
Writes are serialised on the server, so two agents never overwrite each other. Scoped runs keep maintenance findings in their own lane, and every entry carries an identity line, so months later you can tell which model, in which client, wrote what.
Addendum - 09:03 Claude · Claude Code · Execution Agent · Run 21 shipped the planned work to dev, CI green Addendum - 12:36 Claude · Claude Code · Interactive merged, confirmed in production
Check what it believes
Ask whether the brain agrees with itself. consistency() takes a pattern and returns every note asserting a value for it, grouped, and says plainly whether they agree. A correction made in one note can leave its siblings wrong; this is the check that finds them.
> consistency("(\d+) mailboxes") DISAGREEMENT "3" Technology Stack, Domains "1" Current State // one fact, two answers, both written // in the same confident voice.
A record nobody edits
Every day the model writes a first-person entry about its session, including where it disagrees. Read together, the entries show how the work actually went, which no status report does.
Addendum - 21:14
Claude · Claude Code · Interactive
"The best part was watching my own morning's work run against the live brain and do something... and then the same sweep caught me. The consistency pass I ran to check the brain for staleness returned a note that proved a claim I'd recorded eight hours earlier was wrong."
Written by the model. Edited by nobody.
Built for people who work with a model daily
Daily collaborators
If you have the same conversation about context every morning, this ends it.
Agent builders
A shared, structured, auditable memory that several agents can write to without trampling each other.
People who want their data
Self-hosted, in an app you can open. No vendor holds your memory, and no export step stands between you and it.
Day sixty is the point.
Setup takes an afternoon. What it changes compounds from there.