Vocabulary that actually sticks. Spaced repetition, shaped around you.

Building a deck by hand is where most people quit. Three hours of typing words in and you've studied none of them. So rikai ships 40+ ready-made decks: the whole JLPT ladder, all of HSK, SAT, TOEFL, and more, each one tap away. The science here is boring and true. The best study session is the one you actually start, so killing the setup cost is what gets you showing up every day. And daily reps are the entire game.

Point your camera at a textbook page, a handout, a list you scribbled at the back of class, and rikai's AI turns it into a clean deck in seconds. Prefer a spreadsheet? Drop in a CSV just as fast. This matters more than it looks. The words you actually meet in your own materials are the ones worth drilling, and personally relevant vocabulary gets encoded deeper and recalled better than a generic frequency list. You study what's in front of you, not someone else's word dump.

Isolated words are the hardest thing to hold onto. So rikai writes every example sentence out of the other words already in your deck. Review one card and you quietly re-meet two or three more in fresh context. Two effects stack here. Contextual encoding gives each word more hooks to be pulled back by, and every incidental re-exposure works like a low-effort retrieval rep — one of the more reliable ways to make new vocabulary stick. Your deck stops being a flat list and starts reinforcing itself with every session.

If you already speak another language, that's leverage, not clutter. rikai defines each new word in your language and lets one or more languages you know ride along on the same card. When a Japanese word maps cleanly onto a Mandarin one, or an English word turns out to be a loanword in Japanese, seeing them side by side turns a cold new word into a short hop from something you already own. Cross-language transfer and cognate mapping are well documented, and for multilinguals and heritage speakers they're a real shortcut in.

You know the loop. You hit a word, stop, go ask a chatbot what it really means, and lose your place. rikai puts that on the card itself: the definition, the nuance, the way it's actually used, plus example sentences drawn from your own deck. Everything you need to grade yourself honestly is right there, at the exact moment spaced repetition surfaces the word. That timing is the whole point. Meeting a word just as you're about to forget it, and pulling it back from memory, is what makes it stick.

This is the engine. Short 20-word sessions so you actually finish, an algorithm that spaces each word to come back right before you'd lose it, and grading you do yourself, from memory, out loud. That last part carries the most weight. Pulling a word back from a blank beats re-reading it, and spacing your attempts beats cramming them, two of the most replicated findings in all of learning science. Everything else in rikai just feeds this loop. The full citations live in THE RESEARCH.