Cockpit
What to do, in what order, and how long it will take. The evening is decided before you sit down.
We are the brain behind
your AI Tutor
Scroll to explore
You lose time and money with traditional learning methods
Hire a tutor
CHF 60per hour
Finally someone who sees what you’re missing — hour by paid hour. What it costs is everything in between: you’re on your own until the next session, and part of that one goes to catching them up on where you are.
Try to teach yourself
Structural chaos
Free, flexible, and entirely up to you. What it costs is time — the evenings that go into deciding what to work on, and the weeks that pass before you find out something never stuck.
Do it with AI
Limited context
Instant answers, any subject, at two in the morning. What it costs is continuity — every chat starts from nothing, so you explain yourself before you learn anything, and no one notices the topic you quietly stopped opening.
Scroll to see the Unitorly method
Launch in April 2027
Unitorly knows your level of learning…
Financing
3 entriesAccounting · mid-term · in 19 daysAccounting › Balance sheet
11 %
θ (raw)−0.50
θ effective−0.40
Attempts1
Stability2 days
Last practised23 Aug 2026
Review due25 Aug 2026
Last correctness—
Measurement certaintymoderate (±0.80)
Bloom levelApply
Strengths
- Sorts the two sides of the balance sheet correctly
Weaknesses
- Books an increase in capital on the wrong side
- Mixes up equity and long-term debt
New learning session
- Read a balance sheet38 %
- Discuss everyday and economic topics (B1→B2)2 %
- Describe matter cycles3 %
Recommendation from your last session (2026-07-25): Next time, touch on the power and logarithm rules only briefly and spend the time on applying them instead.
…and is able to tell you what to do to achieve your goals.
Give the start prompt to an AI of your choice to make it your private teacher who already knows what to do.
Inside the Unitorly algorithm
This is one competency from the window above, run forward by the same engine that wrote your prompt. Where the curve meets the target, a review is due; after it, the curve starts again and falls more slowly. Pull a slider. Everything you see is computed, not drawn.
- θ effectiveθ after decay — what you would score today, not on your best day. θ × (0.4 + 0.6 × recall).
- 0.85
- Recall todayThe odds you get it right this minute, read straight off the curve.
- 98 %
- Review dueThe day the curve meets the target. That is when it comes back.
- in 8 d
- ProgressHow close to mastered, out of θ and its uncertainty. A wide band keeps it low even at high θ.
- 38 %
- GateOpens at θ 0.5, once at least two answers exist. Until then there is too little to plan with.
- open
- Exam inDays until the exam. Inside the last 42 days the target rises from 90 % to 94 %, and the last review lands the day before at the latest.
- 19 d
The review on day 8 lands at 94 % recall, scores rating 3 — stability 21 → 50.8 d, θ 0.86 → 0.83, uncertainty ±0.50 → ±0.49. The next one falls 10 days later.
Every number here is read out of the engine that runs it. None of it was written for this page.
This and many more features for your organisation.
Leaded studying with AI
Unitorly writes the hour, hands it to an AI of your choice, and reads the report back into your record. Every session starts where the last one ended.
Flashcards
The scheduler matches Anki’s, and it knows your exam date. Nothing you learned quietly slips away.
Knowledge database
One tree for everything you study, at the level you work at. You can see where you stand.
Exams
The exam stops being the thing you avoid. Put in the date and the goals; the plan follows.
Homework
Everything due, with the time it will take. You find out on Monday, not on Thursday at eleven.
Lesson notes
Notes live in the lesson they came from. Months later you still know where to look.
Time table
Enter your week once. After that the app knows which days are full before you do.
Podcasts
Your own material, read out loud. Twenty minutes on the bus turn into twenty minutes of revision.
Analytics
What you actually know, per topic, over time. On the flat days, here is the proof.
Rewards
Points that buy real things, not badges. The hours end up worth something outside the app.
iOS app
The same record in your pocket, offline. Five minutes in a queue are five minutes of work.
We built this for ourselves first.
Unitorly began as a memory. We wanted one place that knew what still needed learning and how well each part actually sat, and that handed exactly that to the AI prompts we were already writing every evening. It was not a product. It was a workaround.
We have worked on it since May 2026, seven days a week, and somewhere in there it stopped being a workaround. It got good enough that keeping it to ourselves made no sense.
Four things still stand between here and April 2027. Setting up an account should take minutes. The rewards have to buy something real, every part of the app needs a guide, and the design is not finished.
Phase 1 · Launch, April 2027
The product
Everything on this page, finished and open to anyone who studies.
Phase 2
We listen
The first users will find what breaks and what is missing. That goes back into the app before anything new gets built.
Phase 3
Our own AI
A chat inside Unitorly that already knows where you stand, so you stop explaining yourself at the start of every session.
Phase 4
A learning management system (LMS)
Schools working inside it, not just learners. That is where this is going.
Early access
Two free months of Premium for going first
Access opens group by group in the months before the April 2027 launch, and the free months start the day you get in. In return, one thing: tell us what annoyed you. Some of it will.