Equathora - Interactive Math & Gamified Learning Platform

An intelligent math practice platform featuring step-by-step guided learning, native LaTeX rendering, and (Gemini) Sigma AI tutoring designed to build true mathematical understanding.

Comment

Inspiration Math education often feels like hitting a wall. Most online platforms rely on static video lectures or multiple-choice questions where guessing is easier than actually understanding. When students get stuck on complex algebraic expressions or calculus problems, they are usually handed instant final answers without understanding where their logic broke down.

We built Equathora to fix this. We wanted to create an interactive, AI-native math workspace that acts like a real tutor guiding learners through step-by-step reasoning, displaying clean mathematical notation, and building true problem-solving confidence.

How We Built It Equathora is engineered as a modern, practice-first web platform designed for continuous learning and intelligent guidance: Frontend & Math Rendering: Built using React, Vite, and TailwindCSS. We created a custom MathText component that dynamically parses raw input into clean, readable math notation across practice problem sets and live chats. Backend & Infrastructure: Powered by Supabase for secure user authentication, progress tracking, and database management, hosted and deployed on Google Cloud infrastructure. AI Tutor (Sigma AI): Powered by the Gemini API on Google Cloud. Sigma AI analyzes intermediate student steps, evaluates logic, and provides targeted hints rather than spoiling solutions. Structured Learning: Designed Learning Journeys and Daily Track missions featuring XP, progress metrics, and diagnostic quizzes to personalize every student's path.

Challenges We Faced Pedagogical AI Tuning: Training Sigma AI to avoid spoiling final answers was tough. We needed Gemini to evaluate intermediate mathematical steps such as verifying if a student correctly applied the quadratic formula x = (-b ± √(b² - 4ac)) / 2a and output supportive, logical nudges instead of raw solutions. Real-Time Text Parsing: Formatting unstructured text outputs from an AI into valid, flicker-free mathematical notation required building a robust client-side parsing pipeline for complex fractions, exponents, and symbols. Designing for Consistency: Balancing complex features like past submission histories and step-by-step diagnostic workflows with a clean, fast interface optimized for daily study sessions.

What We Learned AI as an Educational Scaffolding Tool: AI creates the highest impact in learning when it acts as a mentor that guides reasoning rather than an automated answer generator. The Power of Clear Feedback: Small details matter instant step verification, clear streak metrics, and readable formatting significantly increase student engagement and retention. Building AI-Native Architectures: Integrating Google Cloud and Gemini allowed us to deliver fast, reliable, and scalable AI interactions natively within the student workspace.

Built With

Share this project:

Updates