About JLS

Japanese Language Studies is a structured curriculum for serious learners. It runs from absolute beginner through upper-intermediate, with grammar, vocabulary, kanji, reading, and spaced-repetition practice progressing together in every unit.

Audio is generated with VOICEVOX, an open-source Japanese text-to-speech engine, so every example sentence has a consistent voice you can listen to while you study. The project deliberately uses its own level system (Starter, Foundation, Progression, Mastery) rather than JLPT labels - the two don't map cleanly, and the goal is fluency in everyday and professional Japanese, not test prep.

Why I built this

I moved to Japan and wanted to study the way a good textbook teaches: grammar, vocabulary, kanji, and reading building on each other in a deliberate order, rather than four separate apps that never quite line up.

What I could not find was something that kept going. Most resources are built for the first year and thin out after it, which is exactly when the plateau arrives. You know enough to get by, the easy gains are behind you, and progress becomes hard to see.

So this is built for the whole arc: from your first hiragana through upper-intermediate, each unit adding grammar, vocabulary, kanji, and a reading passage that uses them together. Spaced repetition runs underneath, so the early material is still with you when you reach the hard part.

Credits

Pronunciation audio is synthesized with VOICEVOX (voicevox.hiroshiba.jp); voices: VOICEVOX:猫使アル, VOICEVOX:春日部つむぎ, VOICEVOX:九州そら, VOICEVOX:冥鳴ひまり, VOICEVOX:波音リツ, VOICEVOX:玄野武宏, VOICEVOX:剣崎雌雄, VOICEVOX:雀松朱司, VOICEVOX:WhiteCUL, VOICEVOX:白上虎太郎.

Kanji stroke-order data derives from KanjiVG (kanjivg.tagaini.net), Copyright Ulrich Apel, used under the Creative Commons Attribution-ShareAlike 3.0 license. Our derived stroke dataset is published under the same license at github.com/marcus-saito/jls-strokes.

Kana stroke-order data derives from strokesvg and the Klee One font, Copyright The Klee Project Authors, used under the SIL Open Font License 1.1.

Kanji dictionary data and name readings derive from KANJIDIC2, Copyright Electronic Dictionary Research and Development Group, used under the Creative Commons Attribution-ShareAlike 4.0 license. The derived lookup dataset is published at github.com/marcus-saito/jls-kanjidic2-lookup. KANJIDIC2 and KRADFILE-U were also used to verify stroke counts, grades, and radical information throughout the Reference Library.

Word dictionary data derives from JMdict, Copyright Electronic Dictionary Research and Development Group, used under the Creative Commons Attribution-ShareAlike 4.0 license. The derived lookup dataset is published at github.com/marcus-saito/jls-jmdict-lookup.