PLATFORM FOR EDUCATORS

Turnyourmethodology
into an AI tutor

Upload tasks, solution steps, and hints — our engine guides students through your methodology autonomously. With analytics. With real comprehension measurement.

Book a DemoWatch Video ▶
AI TUTOR · FANO CONDITION
Let's work through the Fano condition. Start simple: imagine a wire between two houses. You plug it in — your friend gets a shock. That's a one. You don't plug in — that's a zero.
Codes A=00, B=01, C=10, D=11. DBAC = 11 01 00 10 = 8 digits
Correct! 8 digits. Now imagine: you want to save space. What if frequent letters got shorter codes?
Type your answer...
33 script elements ⟐ 6 hint levels ⟐ Bayesian student model ⟐ Transfer + Retention ⟐ L2T Protocol (Apache 2.0) ⟐ POMDP formalization ⟐ Not a prompt on top of an LLM ⟐ 33 script elements ⟐ 6 hint levels ⟐ Bayesian student model ⟐ Transfer + Retention ⟐ L2T Protocol (Apache 2.0) ⟐ POMDP formalization ⟐ Not a prompt on top of an LLM ⟐

Seeitinaction

Two short videos: how a student learns and how an instructor creates a course

1 min · A student solves a problem with an AI tutor

STUDENT

2 min · An instructor builds a course in the editor

INSTRUCTOR

Fromideatoaworkingcourse—4steps

01

Watch the demo

A real lesson on video: a student solving a problem with an AI tutor. The tutor guides without giving the answer.

02

Schedule a presentation

Tell us about yourself: what you teach, your methodology, how many students. We'll reach out within 24 hours.

03

Try it yourself

Get trial access. Go through a lesson as a student — see the tutor, hints, and checks from the inside.

04

We'll build your course

Send us your methodology — we'll help create your first course with the AI engine. In your teaching style.

Start with Step 1 →

Not a prompt on top of an LLM.
Custom architecture.

Formalizing education as a control problem with a Bayesian student model

ARCHITECTURE

Knowledge Graph

Many-to-many links between skills, tasks, and errors. Not a flat list — a compositional subject structure.

ENGINE

Bayesian Student Model

Updated on every observation. Reliability scoring: the system is not fooled by cheating.

PLAYER

Methodology Player

State machine: task → plan → step → check → hint → exercise. 6 hint levels without spoilers.

METRICS

Outcome Measurement

Transfer (new task, same principle) + Retention (check after 1–7–14 days). Not 'solved now' but 'understood and remembered'.

Your expertise ×
our engine

Your teaching style

The tutor follows your methodology and speaks in your voice. Students learn 'with you'.

Analytics

Which hints work. Where students get stuck. Mastery growth, transfer, retention — in numbers.

Experiments

A/B comparison of methodology variants on real students. Objective data instead of 'I think this works'.

Open format

Your methodology in an open standard (L2T Protocol). No vendor lock-in. Data can be exported.

Anti-cheating

Reliability scoring: the system knows when a student is cheating. Honest comprehension metrics.

Quality control

You approve every change. Candidate → Validation → Approval → Rollback.

★ "My linear algebra course now runs 24/7" — Professor, MIT ★ "Finally objective metrics instead of intuition" — Methodologist, TU Berlin ★ "Students get hints at my level" — Tutor, 15 years of experience ★ "My linear algebra course now runs 24/7" — Professor, MIT ★ "Finally objective metrics instead of intuition" — Methodologist, TU Berlin ★ "Students get hints at my level" — Tutor, 15 years of experience

Twoaudiences.Oneplatform.

Educators and tutors

  • Scale your methodology to hundreds of students without losing quality
  • AI delivers your steps, hints, and checks — 24/7
  • Analytics: which hints work, where students get stuck, what to improve
  • Students learn 'with you' — the tutor follows your style

Researchers and methodologists

  • Open protocol (L2T Protocol) for formalizing methodologies
  • Structured logs: every step, hint, and error — recorded
  • Objective methodology comparison via transfer and retention metrics
  • Research paper with full formalization (POMDP, hypergraph, student model)

Whatsetsusapart

Typical AI TutorClosed Adaptive SystemLearnToThink
CoreLLM + promptClosed systemCustom architecture
Student modelNoneClosedBayesian, open
ContentOwn catalog onlyOwn onlyAny educator
HintsLLM generatesSystem selects6 levels, designed by methodist
Methodology comparisonImpossibleImpossibleA/B on standard metrics
Open protocolNoNoL2T Protocol (Apache 2.0)
Transfer/RetentionNoNoBuilt-in

How a real methodology works

The 'Fano Condition' methodology: from zero to exam-level problems in 100 minutes

1Increment 1

Wire and current → uniform code

  • Wire metaphor
  • What is a bit
  • Uniform code
  • Task: encode DBAC
2Increment 2

Variable-length code and the problem

  • Why save space
  • Decoding problem
  • Task: find ambiguity
3Increment 3

Fano condition

  • Rule formulation
  • Code tree
  • Task: verify a code
4Increment 4

Building codes

  • Tree-based algorithm
  • Optimality
  • Task: build a Fano code
5Increment 5

Target tasks

  • Exam problem #1
  • Exam problem #2
  • Exam problem #3
  • Transfer task
0
theory blocks
0
practice tasks
0
target tasks
0
script elements
~0
minutes from zero

See the system in action
and ask the team questions

Tell us about yourself — we'll reach out, show you the platform, and help digitize one lesson from your methodology.

1Fill out the form — we'll respond within 24 hours
2A short call — we'll show the system and answer questions
3Trial access + help building your first lesson

Scientific foundation and open protocol

Everything is open: formalization, specifications, data schemas.