←Back

01 / Mathsy · AI education

Mathsy / AI education

Building a better way through difficult maths.

Mathsy brings structured curriculum content, intelligent practice and AI assistance into one learning environment for GCSE, A-Level and IB Maths.

RoleCo-founder / Engineer
ScopeProduct · AI · Web · Content systems
Visit Mathsy ↗

00 / The premise

Not another answer machine. A learning system with context.

  1. 01Structured first

    Curriculum gives every explanation and practice question a place.

  2. 02Help, not answers

    Assistance moves the learner forward one useful step.

  3. 03One environment

    Learn, practise, plan and ask without losing context between tools.

01 / The product

Curriculum structure and intelligent help, in the same place.

One place to learn, practise and ask.

  • Textbooks

    Clear explanations mapped to the full course.

  • Practice

    Questions sit beside the theory they reinforce.

  • AI tutor

    Help arrives with the lesson context already attached.

Content generator

3,302 creditsN

AI content studio

What should we make?

Turn a teaching idea into a complete, editable classroom resource.

Recent resources

▱Quadratic graphs10 questions · Yesterday

▱Ratio and proportion14 questions · 26 Aug

Product demo / Worksheet generatorPrompt → Resource
02 / Mathsy in motion

Short explanations became both a learning format and a distribution system.

115K+TikTok followers

Difficult ideas, made watchable.

Motion lets a proof unfold, a graph move and an abstract idea become visible. The same system that teaches inside Mathsy also reaches learners where they already spend time.

01Normal distribution
02Central limit theorem
03Partial fractions
04Polynomial division
03 / System card

A compact record of what the project made, taught and runs on.

Project system / 01

Mathsy

Active

Achievements

115K+TikTok followers

200+paying subscribers

6,429questions in the curriculum

What we learned

  1. AI becomes useful in education when it understands the curriculum, the learner and the next useful step—not just the question.
  2. Distribution is part of the product. Short-form explanations became a repeatable learning and acquisition loop.
  3. Trust lives in the details: clear notation, visible working and predictable feedback matter more than novelty.

Tech stack

  • Next.js
  • React
  • TypeScript
  • Supabase
  • KaTeX
  • Sentry
  • PostHog

Project state

Live and still moving. The curriculum holds 6,429 questions across 1,750 explanations, and students have worked through 13,938 of them.

Novica Petrovic / Selected workSystem record · Mathsy