Zanaya · Launching summer 2026

The math tutor of
the 21st century.
Personal for every child.

Zanaya is an AI math tutor that sees exactly where a student fell behind - at the level of an atomized skill - and builds the precise learning path to fill the gap. 20 problems, fully understood. Not 100, half-forgotten.

Get early access → How it works
EXERCISE · 042 / LINEAR EQUATIONS
PROBLEM
Solve for x: 3(x + 2) − 4 = x + 6
✓
→ 3x + 6 − 4 = x + 6 distribute the 3 inside
✓
→ 3x + 2 = x + 6 +6 + (−4) = +2, not −6
Zanaya listening…
Gap detected
Mastery: 84%
1,000+
students taught at Zanaya Academy over 5 years
2,000+
students enrolled in summer 2026 program
5 yrs
teaching experience built into the model
700+
atomized skills in our skill tree
The problem

A single grade
tells you nothing.

A caring parent gets the report card: a C in geometry. Now what? They don't know where the actual problem is. The kid doesn't know what to fix. The teacher has no time to diagnose 28 students individually.

"
My child is getting low grades. I'm a caring parent, but I can't figure out where exactly the problem is. Can you help?
- a question we hear from parents daily
01

The system is from the 19th century

One teacher, one large classroom, one grade for everything. Personalized attention is impossible. Real gaps go undetected and accumulate over years.

02

Today's AI tutors miss the point

They're chatbots that assume the student knows what to ask. But if the student knew, they wouldn't need a tutor. That's why nobody actually learns from them.

03

Private tutors are expensive

A good private math tutor in any major city is $20–60 an hour. Few families can afford four sessions a week. The rest go without proper support.

04

Math is learned through practice

Not by watching videos. Not by chatting with a bot. By solving the right problems in the right order. That's what the market is missing today.

The solution

20 problems, fully understood.
Not 100, half-forgotten.

Zanaya diagnoses what a student knows on every problem. It analyzes every step of their handwritten work. It finds the precise skill that's missing. Then it gives the next right exercise.

01 / DIAGNOSE

Atomized skill model

K-7 math broken down into 700+ atomized skills. Each linked to its prerequisites. Each measurable. Each backed by thousands of practice exercises.

02 / ANALYZE

Step-by-step handwriting analysis

Students solve with pen and paper, on iPad, with a smart pen, or by typing. The AI reads every step, not just the final answer. Mistakes get caught where they actually happen.

03 / PATH

Dynamic learning plan

When a gap is found, we route to the exact skill that caused it, even if it's from a previous grade. We patch the gap without losing progress on the current topic.

04 / PROFILE

Live profile of what they know

Every solved exercise updates the student's skill profile. We see what's mastered, what's forgotten, what needs review. Parents get a real picture, not just a grade.

05 / RETAIN

Spaced repetition

A skill isn't mastered when a kid gets it once. It's mastered when they keep it. The system tests it again days and weeks later, ensuring durable knowledge, not crammed facts.

06 / PLAY

Gamified by design

Defending skills as levels. Points, rewards, leaderboards. Progress is visible and earned. Kids don't learn because they have to, they learn because they want to.

Our innovation

The skill tree, alive.

Two views of the same model. On top, the full skill graph - 700+ atomized skills clustered by domain, layered by grade, cross-linked by prerequisite. Below, what diagnosis looks like in real time: the precise atom that broke under that linear equation, traced back through every prereq it depends on.

Note: the trees below are a simplified demo for illustration, not the live production tree. The real model holds 700+ skills with thousands of prerequisite edges.

The skill graph

Every skill knows what requires it and what follows from it. Cross-domain edges (dashed) connect skills that share underlying ideas - fractions to algebra, signs to inequalities.

Mastery profile · live
mastered in progress cross-domain link 5 layers · 34 nodes shown · 700+ in production

Diagnosis: "5 + 12 = 7"

A single mistake on a grade 7 linear equation cascades down. Zanaya finds the broken atom - integer addition - and traces every prereq the student needs to solidify before resuming. The route home, lit up.

Gap found · grade 5 · integers
remediation route broken atom weak prereq 3 prereqs surfaced · est. 4 sessions to recover
Why we're different

We tested
20+ competitors.

We spent months manually testing every AI math app on the market. None of them solves the actual problem. Each one is a chatbot, a calculator, or a list of exercises. Zanaya is a different kind of product.

Capability
Other AI math apps
Zanaya
Knows what you should study
No. Waits for you to ask.
Yes. Diagnoses and routes you.
Handwriting analysis
iPad-only or none.
Pen, pencil, iPad, smart pen.
Catches mistakes step by step
Final answer only.
Every step of the solution.
Routes back to the missing skill
No. Just gives a new answer.
Yes - even from prior grades.
Live skill profile
A topic checklist at best.
A 700+ skill tree, alive.
Curriculum coverage
Mostly US Common Core.
Built ground-up for our market.
Spaced repetition
Rare. No real cognitive science.
Yes. Built into the system.
The team

Mathematicians.
Not marketers.

The three of us are international math and physics medalists, NASA Ames contest winners, and founders of businesses that already work. We're not a marketing agency building another AI chatbot. We're solving a problem we've faced every day for the last 5 years.

Aleksandar
Co-founder · Skills model

Designs the skills model that powers Zanaya - the framework that breaks mathematics into atomized skills and turns student work into measurable progress. Silver medalist at the International Experimental Physics Olympiad and First Prize winner at NASA Ames (UniSpace, 2018). Previously founded ACTIVE СМУТИ, a consumer brand sold across 350+ retail locations in Bulgaria.

NASA Ames 1st Phys. Olympiad Silver CERN Programme Founder, ACTIVE
Simeon
Co-founder · Curriculum

Leads curriculum design and teacher training. BSc in Mathematics and Computer Science from FMI, Sofia University, and First Prize winner at NASA Ames (2018). Created MatheMagical Games, Bulgaria's first gamified math competition, and has personally taught 650+ students over five years - the foundation Zanaya's pedagogy is built on.

FMI Sofia NASA Ames 1st 650+ students taught Creator, MatheMagical
Jacob
Co-founder · Engineering

Builds the AI systems behind Zanaya's personalized learning paths. MSc in Artificial Intelligence at UCL and BSc in Mathematics at Warwick, with a deep-learning thesis on bird-song recognition acquired by a Brazilian research firm. Solo founder of JustCommit, a habit app with thousands of active users, and a two-time Kaggle silver medalist.

MSc AI, UCL BSc Math, Warwick 2× Kaggle Silver Founder, JustCommit

Summer 2026 - opening to the first 2,000 families.

If you're a parent and you want your child to be among the first to use Zanaya, drop your email. We'll reach out before the public launch.

Got it. We'll be in touch soon.