Learnify

Introduction

Learnify started as a question: what would a study platform look like if it organized knowledge by its prerequisite dependencies rather than presenting it as a static, linear list? The product is an AI-powered smart study orchestrator featuring an automated Unlocking Engine, spaced repetition via the SM-2 algorithm, and interactive knowledge graph visualization.

The problem with static learning systems

Existing learning management systems treat education as a flat checklist and treat all topics as isolated silos. They cannot dynamically lock or unlock downstream concepts based on actual mastery, and they have no mathematical concept of cognitive decay over time. For a serious student, that is the inverse of useful.

A study system should know the architecture of the subject — what foundational knowledge is required before advancing, and exactly when a previously learned concept needs review. Anything less is just a digital filing cabinet.

Architecture

I engineered an Unlocking Engine built on top of Directed Acyclic Graphs (DAGs) to govern topic progression. Topics act as nodes, and prerequisites are the edges connecting them; the engine computes prerequisite mastery in the background, autonomously unlocking downstream topics only when all parent nodes are mathematically proven to be retained.

Persistent memory retention is driven by a native implementation of the SuperMemo-2 (SM-2) algorithm. Every user review mutates the topic's interval, repetition count, and easiness factor (EF), automatically recalculating the next optimal review date. On top of this sits an AI generation pipeline utilizing OpenRouter, which parses a single subject prompt and deterministically outputs a strict JSON topological roadmap, instantly seeding the database with topics and dependencies.