How it works
The evidence behind Jukuna
Jukuna isn't a quiz app with AI bolted on. Each part of the product is built on a method that the research literature backs for actually raising exam scores. Here is what we use, why it works, and how you experience it.
Retrieval practice (the testing effect)
Highest confidenceActively recalling an answer strengthens memory far more than re-reading. One of the most replicated findings in learning science.
In Jukuna: Every Jukuna session is recall, not review — you answer fresh questions, then see the worked explanation.
Spaced / distributed practice
Highest confidenceThe same study time spread across days beats one massed session. Spacing slows forgetting.
In Jukuna: A spaced-repetition (SM-2) schedule resurfaces topics you're about to forget, and the session-end card shows what's due tomorrow.
Adaptive / intelligent tutoring
Highest confidenceSystems that adjust difficulty to the individual learner consistently outperform one-size-fits-all material.
In Jukuna: An ELO rating per topic tunes each question to your reach — hard enough to stretch, not so hard you stall.
Immediate, specific feedback
StrongFeedback that explains why an answer is right or wrong, given promptly, accelerates learning.
In Jukuna: Answers are scored instantly with an explanation and an error category, so you learn from each mistake before the next question.
Interleaving
StrongMixing related topics, rather than blocking one at a time, improves the ability to choose the right method on a real exam.
In Jukuna: Sessions interleave sub-topics within a subject instead of drilling one skill in isolation.
Worked examples (with fading)
StrongStudying fully worked solutions helps beginners; the support should fade as expertise grows (the expertise-reversal effect).
In Jukuna: Maths questions scaffold with worked steps that fade as your rating rises, so you're never over- or under-supported.
Self-explanation
StrongExplaining a concept in your own words exposes gaps and deepens understanding.
In Jukuna: After a miss, the "Explain it back" partner invites you to teach the idea back, and a gentle AI judge checks your understanding.
An honest note on AI
Questions, scoring, and explanations are generated by AI and can occasionally be wrong. We treat them as a study aid, not an answer key — your teacher and textbook are the source of truth. Students can flag any question they think is wrong, and teachers review those reports. We do not use student data to train AI models.