PRCM ASRL fuses SM‑2, FSRS, and IRT scheduling algorithms with Groq‑powered AI analytics to maximize knowledge retention — built on React 19 and TypeScript with zero type errors.
Explore the Platform
Each page reveals a different layer of the platform. Follow the recommended path or jump to any section — your theme preference persists across all pages.
The full platform experience. Real user profile with a GitHub-style 365-day heatmap, weekly performance charts, and the complete settings panel.
A fully working study session with real Japanese vocabulary cards, 3D flip animation, live stats, and the SM‑2 difficulty rating interface.
Comprehensive documentation covering every feature, accessibility option, animation control, and deep explanations of the spaced repetition algorithms.
React 19 + TypeScript internals. Performance benchmarks, algorithm conformance testing, KPI gate results, and the full engineering stack.
What's Inside
Every feature is production-grade — accessibility, performance, and pedagogical science baked in from day one.
Light, dark, and high-contrast modes — all meeting strict contrast requirements for comfortable long-form studying in any environment.
SM‑2 and FSRS for study scheduling. IRT-based adaptive testing terminates early once ability stabilizes — fewer questions, same precision.
365-day activity visualization alongside weekly performance charts, accuracy trends, forgetting curve analysis, and leech detection.
Natural-language performance summaries via Groq SDK. Understand weak spots without reading raw numbers — plain English diagnostics.
Four text-size levels, motion reduction, full keyboard navigation, screen reader optimization, and 12 selectable accent colors.
Class leaderboards, biometric integration toggles, teacher-dashboard privacy controls, and multi-user server deployment support.
The Flow
Four interconnected layers — from raw study sessions to AI-powered insight, gated by measurable KPIs before every release.
SM‑2 and FSRS schedulers queue cards at the optimal moment based on measured stability and difficulty parameters.
IRT adaptive engine estimates learner ability and terminates early — fewer questions than fixed-length tests, same precision.
Groq AI explains performance in plain language and surfaces the underlying scheduler curves and forgetting patterns.
KPI gates evaluate accuracy ≥ 75%, retention ≥ 70%, and response time ≤ 6000 ms before any rollout promotion.