Start here

HeroMind is one place where what comes at you becomes action, and where what you decide is kept, so tomorrow, and every AI you use, starts from it instead of from scratch.

It has three layers, and they answer three questions.

AI agents are not a fourth layer. They are the thing that keeps all three current: they file work and sources, read your record before they act, and propose changes you approve. That rule holds everywhere in HeroMind and it has a short name: it advises, you decide.

The loop

Everything in HeroMind runs one loop, whether the input is a morning brain dump, a task you agreed to in a ChatGPT thread, an article you saved, or a finding one of your agents filed overnight.

1. Capture. Put it in, unorganized. Type it, say it, paste it, or let an agent file it.
2. Distill. HeroMind sorts it into the right durable form: a task, a journal line, a list item, a source. You correct in a tap.
3. Approve. Tasks land when you confirm them. Anything about you, a decision, a preference, a standing fact, lands only when you approve it in your review queue, the Desk.
4. Reuse. What you approved is your record. HeroMind reads it before planning your day. Any AI you connect reads it before doing anything for you.

Words this manual uses

• The Desk is your review queue. Anything that would change your record waits there for a yes. The rules that keep it short are in the Library chapter.
• Your record is the context you have confirmed about your work and life: decisions, preferences, beliefs, and your own originals. Each item carries a status and where it came from. It is what agents read about you. The "You" chapter covers it.
• The librarian is the part of HeroMind that compiles sources into wiki pages, writes editorials, builds people pages, and runs every night as scheduled code. It never decides what is true about you.
• Work and Personal are two spaces you switch between in one tap. Tasks, the plan, and notes keep to their space.
• Flag and star. A flag means today: the task is on your focused list until you complete or unflag it, and it rolls forward with you. A star marks a priority on planning views.

Your first week

Day one. Open the Daily Planner and do a brain dump. Tap Process, correct the review list, Apply. Flag two or three things for today. That is the whole ritual; it takes ten minutes.

Day two. Notice that unfinished tasks rolled forward. Decide about each: do it, move it, or drop it. Save one article into the Library from your phone or the browser extension.

Day three. Open the Library. The source you saved has a summary. If it joined a topic with others, a wiki page may already exist. Nothing here needed you.

End of the week. Open You and look at the Desk. If any AI you use has been connected, or if the nightly sweep has run, there will be a few proposed items in your words. Accept what is true. Dismiss the rest; silence is a no, and proposals expire on their own after two weeks.

When you are ready. Connect ChatGPT or Claude (the Agents chapter) so they start from your record. Set up a routine or two. Look at Foundation and Quests only when the daily loop feels like yours; the Guide layer is depth, not a prerequisite.

How this manual is organized

The chapters follow the layers: Do, Know, Guide, then Agents, then the phone and settings. Each chapter says what a screen is for, what the AI does there and what it may not do, and the rules that are not obvious from the interface. The philosophy behind the Guide layer, the ideas from Joe Dispenza, Tony Robbins, James Clear and others that shaped it, is collected in one appendix at the end. It is optional reading. Nothing in Do or Know depends on it.

Updated 2026-09-27 · All chapters · API reference