How it works
Recognising isn't knowing
You can read about closures on a Monday and still fumble the interview question on Thursday. Praxis is built around that gap: it makes you produce answers, explains exactly why you were wrong, connects the idea to its foundations, then pressure-tests it and brings it back until it sticks.
Read the theory in plain language
No jargon-first definitions. Each lesson explains the idea the way you'd explain it to a colleague, then shows it inside real code you'd actually write.
Answer a quiz that explains itself
Every option carries reasoning: why the right answer is right, and why each wrong answer is wrong. Being wrong teaches you something specific instead of just costing a point.
Trace the idea to its roots
Concept genealogy maps each idea onto the one beneath it — useState onto closures, closures onto the lexical environment, that onto scope. This is what turns memorised APIs into understanding.
Get cross-questioned
The mock interviewer doesn't accept a shallow answer and move on. It probes the weakest part of what you said, repeatedly, the way a senior interviewer finds the floor of your knowledge.
Forget nothing
Anything you get wrong — in a quiz or an interview — is scheduled by an SM-2 spaced-repetition algorithm to come back right before you'd forget it. Your weak spots surface on the dashboard.
About the AI features
The tutor and the mock interviewer are the only parts that call a language model, and both are deliberately constrained: they answer only from the lesson content and a grading rubric, so they can't invent facts about the topic. Lessons and quizzes are human-authored, never AI-generated. You can bring your own API key in Settings.