Private AI memory + agents · for founders & teams

Give your company a second brain that actually remembers.

Your company knowledge lives as a private, git-versioned knowledge base, and Claude reads it before it answers — then writes down whatever’s new. No terminal. No config. You talk; it remembers.

Your Second Brain ● remembers

> what did we agree with ‹client› on pricing?

reading the brain…

From your clients/‹client› file and the last meeting debrief: you agreed a fixed €12k build, split in two, plus an optional monthly plan. I’ve logged this question in today’s daily log.

  • clients/‹client›.md
  • meetings/debrief.md
  • daily/today.md
The problem

Your company knows a lot. It just can’t remember any of it.

Everything you’ve learned about your clients, your product, and your decisions is scattered across docs, Slack threads, and CRM notes — and the AI you pay for forgets it all the moment a chat ends.

Scattered knowledge

The context that runs your business lives in a hundred places and mostly in a few people’s heads. Onboarding is slow, decisions get re-litigated, and the moment someone leaves, part of the company walks out with them. There is no single source of truth you can actually ask.

AI with amnesia

You bought Claude or ChatGPT, but it starts every conversation from zero. It doesn’t know your clients, your pricing, or the calls you’ve already made — so you re-explain your business every single time, and nothing it learns survives the next tab. There’s no memory layer underneath it.

What a second brain is

A private AI knowledge base, an operating system for Claude, and a team of agents .

Your company knowledge lives as plain markdown in a git-versioned knowledge base — portable, no vendor lock-in. An operating-system prompt (CLAUDE.md) turns Claude into an assistant that reads the brain before answering, writes new facts down, and keeps itself organized.

01 · The memory

A knowledge base structured around how your business actually thinks — not a generic template.

  • clients/

    Every account’s history, decisions, and psychology in one file.

  • partners/

    Who you work with, the terms, and the state of each relationship.

  • team/

    Your roster, their skills, and the notes you keep on your people.

  • product/

    What you’re building, why, and every call you’ve made along the way.

  • strategy/

    Current goals, live bets, and the reasoning behind each decision.

  • meetings/

    Prep and debrief for every conversation that actually matters.

  • library/

    The articles, docs, and research you’ve imported and digested.

  • daily logs/

    A running journal of what happened, written as you work.

02 · The agents

A team of specialized AI agents that work the brain for you — each with one job it does well.

  • Meeting prep

    Reads the client, the project, and every past note, then hands you talking points, questions, and risks before the call.

  • Meeting debrief

    Turns a raw conversation into decisions, action items, and updated files — nothing gets lost after the call ends.

  • Client files

    Keeps each account profile current: contacts, history, open threads, and what was promised to whom.

  • Partner tracking

    Watches the state of every partner and vendor relationship so nothing goes quiet without you noticing.

  • Ingestion

    Pulls articles and documents into the library — digested and linked to what you already know, not dumped in a folder.

  • Weekly maintenance

    Prunes, links, and reorganizes the memory every week so the brain stays clean instead of rotting.

  • Strategy advisor

    Answers the hard calls grounded in your own past decisions and context — not generic AI advice.

03 · The commands

Plain-language commands anyone on the team can use. No terminal, no syntax to learn.

  • /onboard Teach the brain a new client, partner, or teammate.
  • /remember Write a fact down so it’s there next time you ask.
  • /import Pull an article or document into the library, digested.
  • /brief Get prepped for a meeting or a decision in seconds.
  • /status See where things stand across clients, deals, and projects.
  • /weekly Run memory maintenance and a week-in-review.
Built for non-technical owners

The owner talks; the brain files, links, and remembers. No terminal, no config, no maintenance ritual — the whole point is that a second brain runs itself so you don’t have to think about it.

How the setup works

From your scattered knowledge to a working second brain .

Five steps. We design it around your business, seed it with your real data, install it everywhere you use Claude, teach your team, and tune it for a month.

  1. 1

    Design the memory architecture

    We map how your company actually thinks — clients, partners, team, product, strategy, meetings, library, daily logs — and shape the knowledge base around your business, not a generic template.

    You get

    A memory structure built for how you work, ready for your data to move in.

  2. 2

    Seed it with your real data

    Before day one, we move your real company knowledge in — the accounts, the decisions, the context that already lives in scattered docs, threads, and CRM notes — so the brain is useful the first time you open it.

    You get

    A knowledge base that already knows your business on the day you start using it.

  3. 3

    Install across your Claude surfaces

    We set it up on your machines across Claude Code desktop, Cowork, and claude.ai, wire in the operating-system prompt (CLAUDE.md), and connect the team of agents so everything works from wherever you already talk to Claude.

    You get

    Your second brain live everywhere you use Claude — no terminal, no config to babysit.

  4. 4

    Team working session

    We run a live session with your team and teach them to talk to the brain — the plain-language commands, how to ask, how it remembers — so the owner talks and the brain files, links, and remembers.

    You get

    A team that can actually use it, not a tool that gathers dust after handover.

  5. 5

    30 days of tuning

    For the first month we tune the memory, adjust the agents, and fix the rough edges as your real usage exposes them. Included, not an upsell.

    You get

    A brain that’s settled into how your company really works before you’re on your own.

Who it’s for

Built for founders and teams who’ve outgrown scattered notes .

Founders and teams drowning in scattered knowledge

Teams of 5–50 whose company memory is spread across docs, Slack threads, and CRM notes — and lives mostly in a few people’s heads.

People who bought AI but have no memory layer

You pay for Claude or ChatGPT, but it forgets everything between chats. A second brain is the memory layer that makes it remember your business.

Companies planning AI agents

If you want agents doing real work, they need a knowledge foundation first. Put peanuts in, you get peanuts out — the brain is what you feed them.

Investment

One setup fee. An optional care plan.

You pay once to build and install your second brain. It then runs on your own Claude subscription — no per-seat SaaS fee from us. Keep the optional Brain Ops plan if you want us maintaining it; drop it any time.

Start here Second Brain Setup

€4,900

One-time. Payment can be split.


  • Memory architecture designed around your business
  • A team of specialized AI agents — meeting prep, debrief, client files, partner tracking, ingestion, weekly maintenance, and a strategy advisor
  • Plain-language commands: /onboard, /remember, /import, /brief, /status, /weekly and more
  • Installed across Claude Code desktop, Cowork, and claude.ai
  • Your real company data seeded in before day one
  • A working session to teach your team
  • Role-based access design — private material and sales pipeline kept separate from the team view
  • 30 days of tuning included
Book a call
Brain Ops — optional care plan

€900/mo

Ongoing. Take it or leave it.


  • Memory pruning so the brain stays clean
  • New agents as fresh workflows emerge
  • Ongoing tuning of memory and agents
  • Upgrades as the tooling moves forward
Add Brain Ops

Optional software layer: Routines — a macOS app we build that records your meetings straight into the brain, runs scheduled AI routines on your memory, and keeps everything indexed. Ask about it on the call.

Runs on your own Claude subscription — no per-seat SaaS fee from us, and no data leaving your machines except your own private git backup.

Proof

We didn’t read about this. We run on it.

BinarCode runs on its own second brain every day. Our company memory, our sales pipeline, our client accounts, and our finances are each operated by specialized agents on exactly this architecture — the same knowledge base, the same operating-system prompt, the same command set we set up for you. Concretely: our finance agent parses the monthly bank statements into a live dashboard, our account-manager agents prep and debrief every client call against years of relationship memory, an accountant agent chases and files vendor invoices across four mailboxes every month, and our sales agents keep the pipeline and follow-ups current. We’ve built the same system for clients too, including a Dutch recruitment-technology platform. This is the product we use to run our own company, packaged for yours. Not sure you’re ready for it? Start with an AI Audit — and if you want AI search engines to find you, that’s FOUND.

  • Company memory
  • Sales pipeline
  • Client accounts
  • Finance
  • All agent-run
Common Questions

Second Brain, Answered.

What it is, where your data lives, what it costs, and why the memory layer comes before the agents.

What is an AI second brain?
It’s a private memory layer for your company plus a team of AI agents that runs on top of it. Your knowledge — clients, partners, team, product, strategy, meetings — lives as plain markdown files in a git-versioned knowledge base, and an operating-system prompt (CLAUDE.md) turns Claude into an assistant that reads the brain before it answers, writes new facts down automatically, and keeps itself organized. Unlike a chatbot that forgets every conversation, a second brain remembers your business and gets more useful the more you use it.
Do I need to be technical to use it?
No. It’s designed for non-technical owners. There’s no terminal and no config — you talk, and the brain files, links, and remembers. You ask it questions and give it plain-language commands like /brief before a meeting or /remember to save a fact. We install everything, seed your data, and run a working session so your team can use it from day one.
Where does my data live, and is it private?
On your own machines. Your knowledge base is plain markdown in a git repository you control, and nothing leaves your machines except your own private git backup. It runs on your own Claude subscription, so there’s no per-seat SaaS fee from us and no data flowing to a third-party platform. Because it’s portable markdown in git, there’s no vendor lock-in — the brain is yours to keep.
Does it work with my existing Claude subscription?
Yes — it runs entirely on your own Claude subscription, across Claude Code desktop, Cowork, and claude.ai. We don’t charge a per-seat SaaS fee on top; you pay us once to build and install it, and after that it runs on the Claude plan you already have.
Do all my colleagues need their own Claude account?
No — and this is where most teams overspend. Instead of a seat per person, we set up department-level shared accounts (one for marketing, one for sales, one for operations), so a 12-person company typically runs on two or three accounts. Everyone in a department works against the same shared brain, and role-based access keeps private material and the sales pipeline separated from what the wider team sees. The owner keeps a private account with a personal view.
What do people actually use it for, day to day?
The patterns we see daily on our own brain: walking into a client meeting with a one-page prep built from every past note and open commitment; asking "what do we know about this company?" and getting the whole history in seconds; saying "remember that" after a call and having it filed, linked, and retrievable months later; dropping an article or a PDF in and having it summarized into the right place; a weekly tidy-up that turns raw daily notes into organized long-term memory. Sales pipeline, client accounts, and finance can each get their own agents on top.
Our data is scattered everywhere and half of it is obsolete. How does it get in?
You don’t dump everything in — if you put peanuts in, you get peanuts out. Your team first curates what still matters (we help you decide), then we bulk-index that into the brain during setup so it’s useful from day one. After that, it grows conversationally: new documents go in with a simple /import, articles with a link, and everything you tell it gets filed as you work. Old archives can stay archived.
How is this different from just using ChatGPT or Claude?
ChatGPT and Claude start every conversation from zero — no memory of your clients, your decisions, or how your company works. A second brain is the missing memory layer: it reads your knowledge base before answering, so responses are grounded in your business, and it writes new facts back so the context compounds instead of evaporating when you close the tab.
What does it cost?
Setup is €4,900 one-time, and payment can be split. That covers the memory architecture, the full team of agents, installation across your Claude surfaces, your real data seeded in, a team working session, and 30 days of tuning. There’s an optional Brain Ops care plan at €900/month for memory pruning, new agents as your workflows evolve, ongoing tuning, and upgrades — take it or leave it; the setup stands on its own.
What’s included in setup and how long does it take?
Five steps: we design the memory architecture around your business, seed it with your real data before day one, install it across Claude Code desktop, Cowork, and claude.ai, run a working session to teach your team, and then tune it for 30 days. You get a memory structure, a team of specialized agents, plain-language commands, and a team that knows how to use them.
We already plan to build AI agents — why start here?
Because agents are only as good as the knowledge you feed them. Put peanuts in, you get peanuts out. A second brain is the knowledge foundation agents run on: get the memory layer right first, and every agent you build afterward is grounded in real company context instead of guessing.
Can it capture my meetings automatically?
Yes, optionally. By default you debrief conversationally — "here’s what happened in the meeting" — and the brain files the notes and updates every client and project it touched. If you want meetings captured automatically, we add Routines (getroutines.ai), a macOS app we build: it records your meetings straight into the brain’s memory, runs scheduled AI routines on top of it, and keeps everything indexed. That’s how we run our own setup.
Do we own it, or are we locked in?
You own it. The knowledge base is plain markdown in a git repository on your machines — portable, readable, and yours. There’s no proprietary format and no vendor lock-in. If you ever stop working with us, the brain and everything in it stays with you.

Give your company a memory.

Bring us your scattered docs, threads, and CRM notes. We’ll turn them into a private second brain your whole team can actually ask.

Start a Conversation

Build Your Second Brain