About HeroMind

I started with a better way to plan my day.

I'm Joel Livet. I built HeroMind for myself first, used it every morning for over a year, and kept only the parts that survived daily use. This page is how it grew from a planner into something more connected, and what I think it is for.

The story

Every system got abandoned. This one stuck.

GTD, Things, Todoist, Sunsama, a drawer of dead notepads. I tried them all and quietly left them all. The tools weren't bad. Heavy systems collapse under real life, and a system you stop opening is worse than no system. So I built the lightest thing I could keep showing up to, and let it grow only where daily use asked for it.

Work and life in one day

My work calendar was never going to hold “pick up the prescription” or “practice at golden hour”, and I didn't want it to. But those things compete for the same hours as the Q4 draft. The first version of HeroMind was a planner that put work and personal side by side, each in its own column, without putting either one on the other's calendar.

The parts a task list can't hold

Routines and recurring work kept slipping through, and the useful thing I noticed at 8am was gone by lunch. So recurring tasks, daily routines and daily notes moved in beside the tasks. The morning brain dump became the front door: dump the pile, confirm what becomes a task, a note, or a list item, then pick a focus.

A library I could actually read

For years my reading and research lived in Markdown files, compiled and questioned through Claude Code. It worked and it was ugly. Browsing was a terminal, and nothing I learned ever reached my day. The Library gave that knowledge a web interface: sources kept with their origin, topic pages compile overnight, and editorial pieces that draw connections across sources are proposed for review by default (I can switch a topic to automatic), and answers grounded in what I'd saved.

Agents that needed somewhere to file

Once I had agents reading my email, watching for commitments and researching topics overnight, they needed one place to put what they found. Their tasks now arrive in the Planner with the originating source attached. Their research lands in the Library with their name on it. I set those agents up myself; HeroMind is where their results come back to.

The context that should survive the chat

The most useful thing an AI learned about me kept dying at the end of the conversation. So decisions, preferences and mental models got a home of their own: my record, the ones I choose to carry forward. A connected AI can ask for the relevant part before it starts a new topic, and propose an addition when it notices one. I accept, edit or ignore. When I change my mind, the old entry is superseded, not erased.

Where direction fits

The day was always meant to serve something.

I want my daily plans to connect to my goals, my values and the areas of life I care about. The whole point of an intentional day was that today's actions should connect to where I wanted my life to go. What took time was making that practical without turning it into homework.

The version that works is simple: a major goal gets milestones, I break milestones into tasks, then choose which day to work on each one. Values and life areas sit underneath as context for the choosing, not as a questionnaire you must finish first. Reflection and AI perspectives are optional. To activate a major goal I define a result and a purpose and link at least two routines; that constraint is deliberate, and the rest is there when it helps.

The larger idea

Daily action, accumulated knowledge, your own judgment, and life direction, in one connected product. Run today. Keep what's worth keeping. Decide, and remember that you decided. Point it all somewhere you chose.

That's what I mean by a personal operating system for your life. Not one app that does everything, but the place where the important parts stay connected to each other and to you, while the rest of your work happens wherever it happens best.

Principles

What I hold to, even when a feature would be easier.

You stay in control

AI can interpret, organize, retrieve and propose. Suggested changes to your record, sent as proposals, wait for your yes; routine contributions arrive with their source; and when I say “record this”, it's written now.

Use the best tool for the job

HeroMind doesn't need to own every conversation or workflow. Sunsama is better at timeboxing than I am. Obsidian is a better editor. Keep them, and bring back what matters.

Keep what compounds

Durable knowledge and context should get more useful with time. Working files for one short-lived task can stay with that task.

Today's plan reflects the whole person

Work is part of life, not a separate universe with its own calendar. One day, both columns.

Direction matters

Productivity is only interesting when it serves something you actually value. The day should be able to say what it's for.

Light enough to keep

A system you abandon is worth nothing. Every feature had to survive my own mornings before it stayed.

Who's building it, and how

One person, working alone, using it every day.

There's no team; it's me. I write the code, run the servers, and use the product every morning before I touch anything else. That keeps it honest: if something is annoying, I feel it first. It also means the pace is one person's pace, and I'd rather say so than pretend otherwise.

The beta is private and by invitation. You get the full web app: Planner, Library, your record, and Goals & Direction, free while it's in beta. A narrower Planner-only iOS build runs on TestFlight for people who ask. Connecting your own AI tools is optional and takes a little setup.

What the beta isn't: finished, polished everywhere, or guaranteed to keep every feature in its current form. Things will move. I'll tell you when they do.

Questions, bugs, or a feature you'd fight for: reply to your invitation email and it reaches me directly.

Your data

Brain dumps and decisions are personal. Here is what actually happens to them.

Where it lives

Your data is stored on the hosting platform's managed database, encrypted at rest and in transit. Your data is scoped to your account.

Who sees it for AI features

AI features run server-side; you never paste an API key. Model calls go to OpenAI directly, and through OpenRouter to Google, OpenAI, MiniMax and Anthropic models. I don't train anything on your data.

What connected AI can do

Read what you've made visible to it. Add tasks with a source link and Library sources with attribution, and keep its daily notes separate from yours. Propose additions to your record for your review. It has no delete action.

What you can do

Keep a source or page private, or accept a suggestion privately, and connected AI can't read it. Change your mind and the old entry is superseded, not erased. Revoke any connection from the token page linked in Settings. A connection acts with your access, reading and writing alike; there is no read-only mode yet.

Preview draft. A full privacy policy and terms of use are being written for public launch; until then this section is the reference, and in-app export and account deletion are on the pre-launch list.

Take command of your day.

HeroMind is in private beta. Request access and I'll send an invitation when a seat opens.

Full web appFree during betaAccess by invitation
A narrower Planner-only iOS beta runs on TestFlight; invitations mention how to get it.