Math is learned by doing.
So that’s how we teach it.
Most of the market still teaches math by telling — a lesson, then a video, then maybe some practice. For a lot of learners that order is backwards, and the content behind it is frozen the day it ships. MathPath is the other way round: a problem first, guided thinking, and a library that keeps getting better. We call the goal Blitz Learning — the most effective and most enjoyable way to learn. Here’s the thinking behind it, and how we got here.
Why we exist
The gap we’re closing.
Four things are true at once, and together they leave a real hole:
A chatbot alone isn’t teaching
Raw AI will happily hand over an answer, hallucinate a step, or explain at the wrong level. Generation is not pedagogy, and it’s not safe to put in front of a struggling student unsupervised.
No single team can write it all
Every skill, every misconception, every exam, every language — the space is too large to hand-author and keep current.
Textbooks don’t get updated
They’re correct-ish and frozen. Errors and gaps live for whole editions.
Static practice sites rot too
The “free practice site” model decays the same way — nobody is watching which questions confuse people and fixing them.
So the real need is a system that is question-centric (you learn by solving), actively guided (it coaches your thinking instead of lecturing), and continuously improving (it gets better from use, with both AI and people in the loop). That’s the whole idea.
Our philosophy
Understanding is earned, not handed over.
The goal isn’t “get the answer.” It’s to build the ability to solve — to look at something unfamiliar, figure out what it is, choose an approach, and check your work. A few beliefs shape everything we make:
Productive struggle, not passive reception
We hand you an accessible problem and let you wrestle with it for a moment before any explanation appears. Let the pattern emerge — a lesson handed over up front is a lesson half-skipped.
Nuggets pulled, not pushed
Rather than “study this 12-page lesson first,” you pull a small, just-in-time nugget — a definition, a worked idea — exactly when you feel the need. The references are on tap, not force-fed.
Thinking in three moves
Identify what kind of problem it is, choose a strategy before touching the arithmetic, then calculate. Naming the strategy move out loud is the point: most “I’m bad at math” is really “I didn’t know how to choose an approach.”
Many roads up the mountain
Real math rewards seeing multiple valid methods, not memorizing one. We treat method as a first-class thing — and want learners to share their own route, because seeing a peer’s way is often what unlocks it.
Mistakes are data, not shame
Wrong options are deliberate and mapped to real misconceptions, so a wrong answer produces a specific, useful explanation — not a buzzer.
The engine
A library that improves the more it’s used.
A textbook is correct-ish and frozen; a free practice site rots the same way, because nobody is watching which questions confuse people and fixing them. MathPath is a loop instead — each part doing what it’s best at:
AI drafts
Problems and full walkthroughs, per skill, per difficulty, per language.
People review
Experts approve, edit, or reject — never auto-applied blindly.
You practice
Real use shows where explanations land and where they don’t.
Data improves it
What confuses people becomes the priority queue for the next fix.
And because every change is versioned, a fix carries the memory of what went wrong before — so the same mistake doesn’t creep back. The result compounds: the more it’s used, the better it gets, the more it’s used.
Who we serve
Built for the whole room.
Students
Learn, master, prepare, and track your own progress — the core of everything.
Teachers
Create a class, share a join code, assign problem sets, and see how students do.
Curators
The human half of the loop: review reports, approve edits, and steer content quality.
The focus today is middle school through high school and into college — the segment with the clearest gap and the strongest fit for the method. Younger learners are a deliberate later step, not a now.
Across borders
A good question is an asset in every language.
None of this is specific to one country’s curriculum. Struggle-first solving, the three moves, the improvement loop — they port. So the playbook is repeatable: win one exam in one market, prove the loop, then point it at the next. And questions travel — a strong problem authored in Farsi can be reused in English or Arabic, so every market we enter makes the corpus richer for all of them rather than starting from zero. A working Farsi locale, with a real Konkur archive behind it, already proves this in miniature.
Our story
How we got here.
MathPath didn’t start as an app. It grew out of years of teaching, building, and rebuilding — each step pushing toward the same goal.
We started with pathways
Guided routes through existing math content — and a growing conviction that the path itself, the curriculum, needed rethinking.
We built our own curriculum
So we did the hard thing: an entirely new math curriculum and a fresh approach to math education, designed around doing rather than telling.
Tap-Tap step-by-step solutions
To meet the need then and there, we built guided, tap-through solving — “Tap-Tap.” It worked, but hand-authoring every walkthrough was expensive and slow.
Generative AI changed the math
When capable LLMs arrived, we leveraged them immediately — drafting and enriching content at a scale and speed a human-only pipeline never could. That became the platform you see today.
Idea Pool
Along the way we created Idea Pool — a place for educators and learners to cherry-pick the golden, atomic ideas from books and beyond, fuel for memorable moments in every lesson.
What we’ve built
More than an app.
The platform you’re looking at is the latest piece of a larger body of work — each part feeding the others.
A new math curriculum
A comprehensive rethink of how math is sequenced and taught (yet to be fully published).
Workshops & classes
Run with real students — where we tested and validated the methodology before scaling it.
Idea Pool
A growing library of atomic “golden ideas,” shared across educators and learners.
Tap-Tap solutions
Our step-by-step guided-solving engine — now supercharged by generative AI.
Fully interactive lessons
Mixed-media, modular lessons — interactive modules, visuals, and easy-to-read text — that make concepts click.
Honest status
What’s live today — and what’s next.
Shipped and in use
- Step-by-step solving with identify → strategy → calculate
- Daily review mix, a mistake bank, and a full journey history
- Quiz builder and guided practice (formative or graded)
- Exam prep: SAT, AP and Konkur — diagnostic, study plan, full simulation, score estimate
- An always-available AI tutor that guides without spoiling
- Ask: upload your own question and get a guided walkthrough
- Teacher classes and assignments
- A Farsi locale with cross-language question reuse
Where we’re going
- Learners sharing their own solution methods
- A surfaced challenge tier for the strongest topics
- Richer media — graphs, video, interactive modules
- More languages and more national exams
- Engaging content for younger learners
- Physics Path and Chemistry Path on LearnHub
MathPath is the first path.
LearnHub is a home for learning by doing. MathPath proves the method on math; the same engine — problems first, guided thinking, a corpus that improves with use — is built to carry to other subjects. Physics Path and Chemistry Path are where it goes next.
Start with one problem.
The fastest way to understand MathPath is to try it. Pick a username and solve one.
Free forever to start. No credit card.
Questions, ideas, or partnerships? mohsen.danesh@learnhub.com