Every screen reads the same record.
Rheticus is a desktop app with a small surface. Each screen answers one question, and all of them read and write the same learning record. Every picture here is the real app in the beta, with an example learner's data.
Today: what should I do now?
The next thing to do, sized to the time you have, across every subject, with the reason for each. Your plan for the day and how the week is going sit beside it.
The app as it runs in the beta, with an example learner's data.
- Every suggestion has a reason
- Review overdue by four days. Ready for unseen tasks. Next in the module. If the planner cannot say why, it does not suggest it.
- Fitted to your time
- Say you have 15 minutes or 90, and the plan changes to fit.
- Study days, plainly
- A year of study days and a plain count of the streak. No rewards and no warnings.
The roadmap: where am I on this goal?
A goal mapped by the Curriculum Architect into areas, modules and lessons, with the state of every lesson in view and the reviews that are due.
The app as it runs in the beta, with an example learner's data.
- Counts, never a percentage
- Four retained, one demonstrated, two practised. An average would hide the weak spot you need to see.
- Prerequisites that block
- A lesson stays locked until what it builds on is practised, and the lock names what it waits for.
- Evidence one step away
- Select any lesson to see every attempt behind its state, and its notes and sources.
A lesson: taught before it is tested.
Your Tutor asks where you are starting from, teaches in parts with examples and pictures, answers your questions, and checks only when you say you are ready.
The app as it runs in the beta, with an example learner's data.
- Pictures as you learn
- Plots drawn with Python and diagrams drawn as needed, in the app's own colours.
- Another way in, any time
- Explain it differently, make it simpler, another example, go deeper. Or ask in your own words.
- Three questions to finish
- The check comes last. Two right out of three, and the lesson is understood with help.
Practice: does it stick?
Reviews fall due on a spaced schedule. Practice sets, quizzes, flashcards and the mistakes worth keeping all live here.
The app as it runs in the beta, with an example learner's data.
- Reviews that space out
- Seven days after you demonstrate a lesson, then sixteen, then thirty-five. Miss one and it comes back sooner.
- Marks you can check
- Work marked by a model waits here until you confirm it or overturn it.
- Mistakes kept
- Every miss is kept with the misconception behind it, so you can come back to exactly that.
Notes: where is that thing I learned?
A vault of plain Markdown files on your disk, filed by track and lesson, with your PDFs, books and pages beside them.
The app as it runs in the beta, with an example learner's data.
- Files you own
- Any editor can open them. Mathematics, code and diagrams render in the app.
- Your sources, cited by page
- Bring in PDFs, Word files, slides, books and web pages. Agents teach from them and cite their pages.
- Search by meaning
- With an embedding model, search finds what you meant as well as what you typed.
Schedule: when will I study?
Your week, planned into the hours you said were free, around everything already fixed in it.
The app as it runs in the beta, with an example learner's data.
- Planned, not nagged
- Blocks go where you said you could study. Move them, skip them, or replan the week.
- Ask the Coach
- Whether this week is realistic, and the one change that would make it so.
- Your calendar, if you like
- Connect a calendar and study is planned around your busy times.
Skills: the methods your agents follow.
Fifteen Kircher skills, from mapping a field to making it stick. Each is a plain document you can read, and each can be switched off.
The app as it runs in the beta, with an example learner's data.
- Read every one
- A skill is a SKILL.md file. Nothing about how you are taught is hidden.
- Add your own
- From a file, from GitHub, or written in your own words, for learning in general or for one field.
- Guidance, never power
- Skills shape how an agent works. They never run code or gain tools.
Toolkit: techniques that work, ready to use.
Thirty learning techniques, from explaining it simply to asking what if. Pick one, say what it is for, and the right agent takes it from there.
The app as it runs in the beta, with an example learner's data.
- Slash commands, too
- Type / and a technique's name in any conversation: /whatif, /spot-the-error, /autopsy.
- Tools of your own
- Describe a method you like. The AI drafts it, and you choose which agents use it.
- Lesson maps
- Every topic and subtopic of a lesson, a module or a track, connected as a flow, to see the whole and revise from.
What the beta includes.
All of it, unlocked, while the beta runs. After the beta the free version keeps one track, and nothing fundamental to learning is held back.
Tracks and roadmaps
- A whole field mapped from a goal in your own words
- Skip, refresh, learn or master each area
- Placement tests, checkpoints and projects
- Edit, split, merge and reorder lessons
- Changes proposed with reasons, applied when you accept
Lessons and practice
- Lessons taught in parts, with pictures
- Socratic questioning that asks before it tells
- Code questions marked by running tests on your machine
- Rubrics with named criteria
- Confirm or dispute any model's mark
Mastery and reviews
- Five states, moved only by evidence
- A full evidence log for every lesson
- Misconceptions recorded by name
- Reviews on expanding intervals
- States that go down when a review is missed
Your data
- One SQLite file on your computer
- Notes as plain Markdown files
- Markdown export of everything, at any time
- Scheduled local backups, and restore
- No account, and no reports unless you turn them on
Start keeping the record.
Free during the early beta, with everything unlocked. Join the beta and the Windows build comes to you first.