An adaptive tutor for proof-based mathematics.

Gradus is designed for subjects where learning depends on definitions, theorem hypotheses, proof patterns, and cumulative concept structure.

Concept-level mastery

The engine tracks understanding at the concept level rather than treating every question as an isolated flashcard. Lessons, reviews, and practice sessions can target prerequisites and neighboring ideas.

Deterministic grading

Gradus uses structured answer specifications, exact matching, regex checks, LaTeX-aware normalization, and multiple-choice grading. This keeps feedback predictable and explainable.

Course packages

The app loads course packages with lesson JSON, concept graphs, item families, and optional learning sidecars. Published packages now cover proof-based analysis courses and machine learning prerequisites including linear algebra, convex optimization, and statistical inference.