Catherine Theresia Roeroe

Learning Product · Open Source

moursey.learn

A living harness for truly adaptive learning.

Your own AI tutor that builds a curriculum around you — and changes as you do. Tell it what you want to learn; it maps the path, teaches one step at a time, remembers how you think, and keeps an annotated textbook as you go.

Bring your own model (Claude / GPT) · No sign-up · Your learning stays yours · MIT

Most learning is frozen the day it’s made.

A course is one fixed path for everyone, frozen the day it shipped. Your AI, on its own, is a brilliant but forgetful genius — it waits for you to ask, forgets between chats, and only knows what it knew at training time.

moursey turns that engine into a teacher: it remembers you, holds the plan, keeps things current, and has a method — so it actually takes you somewhere over weeks and months, instead of handing you a great answer you scroll past.

The design

A model of a learner, plus a loop.

moursey isn’t a curriculum. It’s a model of you and a loop that reads that model, teaches the single highest-value next thing, and updates the model from how you engaged. The curriculum is an output — regenerated every session, never a static list.

THE LOOPevery session1Knows you2Picks next3Teaches4Textbook5Adapts
  1. Knows you

    Reads its model of how you think and what you’ve genuinely grasped — not what you claim.

  2. Picks your next thing

    The single highest-value step, inside your zone of proximal development.

  3. Teaches it

    In your best-fit mode, current to the field, inventing a tool when words aren’t enough.

  4. Writes your textbook

    Every idea becomes a page — the field’s words alongside your own notes and analogies.

  5. Understands you better

    Updates the model from what you demonstrated. Then the loop runs again.

Five pieces, working as one.

Like a great tutor, a good textbook, and the right tools — rolled together. New features are these pieces producing something new, not new architecture.

  • The brain

    A model of how you learn

    It tracks what you’ve truly demonstrated — watch it fill in, in color — and uses it to choose your perfect next step. The part that makes it yours.

  • The textbook

    Half the field’s, half yours

    Every idea becomes a page: the canonical explanation paired with your own notes and analogies, so it sticks and you can come back to it.

  • The expert

    Stays current

    It sweeps the live frontier of your field, so what you learn is true now — not a frozen snapshot from training time.

  • The tool-maker

    Builds new ways in

    When words aren’t enough, it generates a fresh explainer, interactive, or exercise on the spot. This is what makes it truly adaptive.

  • The tutor

    Your own AI, running the loop

    Claude or GPT, putting the other pieces in motion each session. You bring the intelligence; the harness brings the structure.

Why not just take a course — or ask your AI directly?

A course or videoJust your AI alonemoursey.learn
Built around youSame for everyoneAnswers what you askLearns how you learn
Knows what to teach nextA fixed pathNothing — you must know what to askMaps the journey, picks your next step
Remembers youNoForgets between chatsNever starts cold
Tracks what you’ve masteredNoNoYes — and aims at your gaps
Keeps you activeYou watchYou read answersYou explain and practice
Something to revisitSlidesScrolls awayYour own growing textbook

The short version: your AI is a brilliant engine — but on its own it’s a forgetful genius who waits for you to ask, and only knows what it knew at training time. moursey turns it into a teacher: it remembers you, holds the plan, keeps things current, and has a method.

From the learning sciences

Built on how people actually learn.

A few non-negotiables, borrowed from the science of learning:

  • Mastery is demonstrated, not claimed. Credentials say where to probe — never what you know.
  • Frame before mechanism. Purpose and big-picture placement come before any detail.
  • Canonical paired with your own encoding. The field’s words and your analogies, side by side.
  • No learner meta-work. It self-improves; you never get homework about the tool.
  • Motivation by autonomy, competence, identity. Not points, streaks, or gamification.
  • Honesty over flattery. The true, useful read — the non-sycophancy it teaches.

Start learning anything.

It can teach any subject — if there’s no ready-made guide for your field, your AI just builds one. You’ll need an AI that can open files; Claude Code (included with a Claude Pro or Max plan) is smoothest.

  1. 1

    Fork it

    Your fork is yours — your map, notes, and progress live as plain files you own.

  2. 2

    Open it with your AI

    Claude Code, a Claude Project, or a ChatGPT custom GPT all work.

  3. 3

    Tell it your goal

    “I want to understand X.” Open index.html to watch your living map grow.

Or clone it

git clone https://github.com/croeroe/moursey.learn.git

MIT licensed · © 2026 Catherine Roeroe · “Learn freely.”