AI that teaches your lesson — and stops there.
Students already use AI. The question a school faces is not whether, but whether the AI they use knows what was taught, refuses to hand over the answer, and costs a predictable amount. A general chatbot fails all three.
The AI your students are already using
Right now, at nine in the evening, a student is asking a general chatbot to do their homework. It will answer, using a method their teacher did not teach, with no record and no limit. That is the status quo you are competing with.
- It explains a different method, so the student is now wrong in a new way.
- It gives the finished answer, which is the one outcome the exercise was designed to prevent.
- The school has no idea it happened, and no way to guide it.
- Where a school does adopt AI, the bill is open-ended and impossible to plan around.
Three tools, all anchored to your content
Each one starts from a lesson your school published, which is the whole difference.
Lesson drafting
Generate a first draft of a lesson from your own unit, in the structure the editor uses. It lands as a draft for the teacher to edit, never as something published.
Exercise generation
Produce a practice set for a lesson, with the question types the platform already marks, so drafting an exercise is minutes rather than an evening.
A voice tutor
A student talks through the lesson out loud and gets it explained back. Useful for languages and for the students who will never type a question.
A chat tutor
The same boundary in writing, with history kept per student and per lesson, so a conversation picks up where it left off instead of restarting cold.
Credits, not a surprise invoice
All three draw on credit pools your school buys. Usage is visible per pool and per branch, and a pool that runs low is a number you can see coming.
Bounded before it spends
Every generation and every tutor session checks the school's balance before the model is called, so an unpaid or empty account costs nothing at all.
AI credits bought against credits used, by pool
Three separate pools — lesson drafting, voice tutoring, chat tutoring — so heavy use of one never quietly drains another.
- Bought
- Used
Illustrative figures — the shape of the report, not anyone's real numbers.
How AI enters a lesson
A teacher is in the loop at the only point where it matters.
- 1
The teacher asks for a draft
From an existing unit, so the generated material follows the syllabus the school actually teaches.
- 2
The teacher edits and publishes
Nothing generated reaches a student until a teacher has read it. This is not a setting — there is no path that skips it.
- 3
The student studies the published lesson
And when they get stuck, the tutor is right there, scoped to that lesson and to nothing else.
- 4
The school sees the cost
Credits spent are attributed per pool and per branch, so AI is a line you can manage rather than a mystery.
Why bounded AI is the harder and better product
Attaching a general chatbot to a school takes an afternoon. Making one that helps without undermining the teacher is where the actual work is, and it is the part most products skip.
Scoped to the published lesson
The tutor answers from the lesson your school wrote. That is why the method it explains is the method the student will be examined on.
It will not finish the work
The tutor exists to unstick a student, not to complete the exercise. Handing over answers would make the exercise data worthless and the teacher blind.
Metered in credits you own
Credits are derived from real model usage and bought as capacity. They accumulate, they never expire, and they are frozen rather than burned if a subscription lapses.
Spending is capped at the door
Rate limits and balance checks sit in front of every model call — and on the paths that spend money they fail closed in production, so an outage can never turn into an unlimited bill.
Questions schools ask about AI
- Will it just do the homework?
- No. The tutors are built to explain the step a student is missing, within the lesson they are on. Producing the finished answer is the outcome they are designed to avoid.
- How is it charged?
- In credits, bought per pool, derived from real usage. They are capacity your school owns — cumulative, never expiring.
- What if we run out of credits?
- The feature stops and tells you, rather than silently degrading or billing on. Your principal tops the pool up and it resumes.
- Can a teacher turn it off?
- AI is a capability your school chooses to buy. A school that buys no AI credits has no AI, and nothing about the rest of the platform changes.
Keep reading
The parts of the platform this page touches.
Give your students AI that agrees with their teacher
Create an account and we will walk you through how the credits work.