I'm Robert Groenendijk, based in Dubai, UAE. I do IT and information management for local government, and I run my own company alongside it. That combination is why MoonBrain exists: two jobs, one set of hours. My Brain is a folder of markdown files in Obsidian with an AI agent that can read and write in it. It runs my mail triage, my task board, my meeting notes and my client reports. Not a chatbot I paste things into and lose again, a system that keeps my context and does the work. Your turn. Four lines is enough: 1. Name and where you are. 2. What you do for work. 3. The task that eats the most of your week. 4. What you would do with those hours back. Done Level 1 already? Post a screenshot of your first note too. Then go to Level 2 and connect your agent.

Everything you install here keeps getting updated. This post explains exactly how, so you know what you are paying for. When: the first working day of every month. Predictable, so you can ignore it the rest of the time. Where the updates come from: I use this system every working day on real client work. When something breaks, when a better route shows up, or when a task keeps needing the same correction, I fix it in my own system first. Those fixes are what you get. This is not content written for a course, it is a system that gets sharpened by use. How to get it: type update my Brain. Your agent fetches the new version, tells you in plain language what changed, and runs any migration for you. About two minutes. What an update never touches: your notes, your cards, your board, your own rules. Only the system folder gets replaced. That is why you never have to check anything before updating. Why monthly and not weekly: updating costs you something every time, however small. Weekly gets ignored after a month or two, and then everyone is running a different version and helping each other becomes impossible. One good release beats four thin ones. Want it sooner? There is an edge channel that moves continuously, for people who like being first and do not mind the occasional rough edge. Say so in Q&A and I will switch you over. Quiet months happen. Some months bring three real improvements, some bring one. I will tell you which it was rather than dress up a small change as a big one. A quiet month means the system is stable, not neglected. Every release gets one post in this category: what changed, and what it actually saves you.

This system grows out of real work. Mine, and from here on yours too. The honest limit: my Brain improves because I use it every working day on client work. Your work is different. I cannot properly test a playbook for a job I do not do. So requests here get built together with the person who asked for them. How to post a request. Answer these four things: What task costs you the most time? How often does it come up? What do you do now, step by step? What should the system hand you when it is done? Why four questions instead of a wish list: a wish is hard to build. A description of your actual steps is already most of a playbook. What happens next. Three people asking for the same thing is a pattern, and patterns get built. One person asking gets a real answer in the thread on how to do it in your own setup, and the request stays on the counter. Most requests turn out to be a variant of something that already exists, and then it is a small change that can be in the next release. The deal: if you ask, you test. I write the first draft, you run it on your real work and tell me what actually happened. Two or three rounds and it is done, and then it goes to everyone in the monthly release. Built something yourself that works? Send it in. If it goes into the core system, your name is on it. Small requests are welcome. They are usually the ones that get built first.