ClassCube

Building Trustworthy
Math Learning Infrastructure.

We combine knowledge maps, question banks, and learning analytics to design and implement math education systems that actually work in the AI era.

Since 2013 EdTech 7 patents GPT · Claude MCP Integration
Mission

Beyond AI Generation —
A Complete Learning System

We don't just generate content — we verify, structure, and connect it to learning states. AI is designed not as a tool that produces answers, but as an engine that connects students to their next learning step.

Verified Questions & Structured Knowledge
We check for errors across answers, explanations, formulas, choices, and context to maintain quality.
AI that Reflects Learning State
We identify where students are and connect them to the next learning they need right now.
Math education without language barriers
We translate verified Korean math content into multiple languages, delivering the same quality to students anywhere in the world.
What We Build

4 Pillars of Math Learning Infrastructure

AI Question Quality Control

AI verification creates hallucination-free math question data. Automatically checks answer, explanation, formula, and choice errors to maintain large-scale question bank quality.

Math Learning Engine

Diagnoses learning status through knowledge map analysis and delivers coaching guides to AI to connect personalized next learning steps.

Question Bank · Authoring

The core asset of math learning infrastructure. A content management system covering storage, classification, search, combination, exam generation, and typesetting.

Translation · Twin Question Agent Beta

An AI agent that translates verified questions into multiple languages and creates twin questions on the same concept, rapidly expanding the global question bank.

Why ClassCube

Accuracy, Consistency,
Verifiability

In educational settings, trustworthy results matter more than fast generation. We build systems that meet that standard.

Knowledge Map + Question Bank + Learning Data — not solved by LLM alone
AI is used as a verifier and assistant, not a generation tool
Manages both question quality and learning flow simultaneously
Implemented as an integrated system, not feature units
IP Portfolio

3 Korea patents, 4 US patents

A domestic and international patent portfolio for our core technologies in question banking, knowledge maps, and learning analytics.

Math Similar-Question Search
KR Registered

Method and implementation for searching documents containing formulas

US Registered

Method and implementation for searching documents containing formulas

Knowledge-Map Weakness Analysis
KR Registered

Knowledge-map node weighting algorithm for learning analytics

US Registered

Knowledge-map node weighting algorithm for learning analytics

Learning Recommendation & Inference Engine
KR Registered

Learning analysis and recommendation method using four knowledge maps

US Registered

Learning analysis and recommendation method using four knowledge maps

US Registered

Deletion and creation of links between knowledge-map nodes for learning recommendation

History

Math education technology built since 2013

2013
ClassCube founded. Math education technology development begins.
2014–16
Mobile learning service and wrong-answer notebook service developed.
2017–18
R&D of similar math problem search, learning analytics, and inference/prediction engines.
2019–22
Expanded question banking, teacher platform, open question API, and LMS development.
2024
Built and supplied HawaiiClass, HawaiiMath, and PlanMath learning platforms.
2025
Advanced AI verification agents — achieved hallucination-free math question data. Integrated knowledge map-based coaching guides.
2026
Official launch of kmath.app. Global service begins as a multilingual math learning platform.

Partnership & Collaboration Inquiry

ClassCube welcomes service integration and technology partnerships with education platforms, publishers, academies and LMS providers, and global partners.

Submit Business Inquiry