A note on the library

Sources & study notes

New to the examination system? Understanding Kanken explains the level map, exam skills and study starting points.

The curriculum

Grades 1–6 follow the 1,026-character elementary allocation in Japan’s Ministry of Education curriculum: 80, 160, 200, 202, 193 and 191 new kanji. These are Japanese school levels, not prerequisites based on a learner’s age.

Grades 7–9 use an EAJ-selected sequence derived from the new-character headings in Mitsumura’s 2025 junior-high lesson worksheets: Year 1, Year 2 and Year 3. Our extraction yields 392, 397 and 241 distinct characters. These are locally audited counts, not publisher-certified totals or official national per-grade lists.

Our Grade 9 collection adds an explicitly selected 80-character completion group to that 241-character core, making 321. It is not attributed to a Mitsumura appendix. Together, Grades 1–9 contain all 2,136 jōyō characters, with each identity introduced once. Example words may include characters from other levels. Kanken Pre-1 and Level 1 add 849 and 2,651 new written forms respectively.

Beyond jōyō: Kanken collections

The advanced inventories were reconciled against the Japan Kanji Aptitude Testing Foundation’s KanjiPedia head entries and level badges. We count new written forms, not every reading or printed variant: Pre-1 has 849 and Level 1 has 2,651 after earlier forms are removed. The already-introduced written form 芸 does not gain another slot for its historical ウン reading. These library totals are not substitutes for the examination’s approximate cumulative character counts.

Pre-1 Parts A–C each contain 283 characters. Level 1 Parts A–E contain 295 each; F–I contain 294 each. Parts are EAJ organizational units, not official exam levels. Level 4, Level 3, Pre-2 and Level 2 may still test new skills after Grade 9 even though they add no new written characters to this inventory.

Our public order starts with editorially selected vocabulary, then groups characters by classical radical and stroke count, with Unicode order breaking ties and selected compound pairs kept together. It is a reference sequence, not a frequency ranking or a proven optimal learning order. The internal corpus-based research sequence is not used for this release.

Advanced cards show dictionary on and kun readings with English character glosses. The opening vocabulary includes checked examples. A dot within a kun reading separates the kanji reading from following okurigana; a hyphen marks a bound reading. Readings vary by word, and these collections do not cover every skill required by Kanken. Rare, historical and name-oriented forms remain available for reference.

Dictionary and example data

Kana and Grades 1–6 were adapted from the Japan Undercover study library. Grades 7–9 add dictionary-backed examples with spelling, reading and sense restrictions checked against JMdict. Frequency-ranked choices were reviewed and selected entries edited for teaching clarity. This is an independent study resource, not a publisher-approved textbook.

Readings, meanings and example-word data draw on KANJIDIC2 and JMdict, published by the Electronic Dictionary Research and Development Group under CC BY-SA 4.0. Adapted lesson datasets are available from each lesson under that license. Grades 7–9 also link entry-level provenance in their downloads.

Examples teach a whole word in context, not every reading of an individual character. Some compounds have irregular whole-word readings. English character glosses are memory aids; they do not define every compound containing that kanji. Some later characters occur mainly in formal, historical or specialist vocabulary. A small number of advanced dictionary gaps use concise EAJ English glosses and readings checked against the linked KanjiPedia head; those entries are identified in the downloadable provenance.

Character forms and stroke guides

KanjiVG by Ulrich Apel and contributors supplies the stroke geometry under CC BY-SA 3.0. Geometry and order are preserved. Animation, display colors and tracing are adaptations under the same license. Attribution and source revision; original SVGs are linked from the viewer.

Our inventory uses 剝, 塡, 頰 and 𠮟 as the jōyō identities. The related forms 剥, 填, 頬 and 叱 do not receive extra character slots. Printed fonts, handwritten forms and dictionary stroke counts can differ. The guide displays the KanjiVG form, including variant treatment in 餌・遜・遡・餅・謎・僅・葛・牙・賭. It is a study aid, not a statement that every examination accepts every glyph variant.

Audio and typography

Hiragana, katakana and Grade 1 offer example-word audio generated with ElevenLabs, Akari / Multilingual v2. It plays only when requested. Grades 2–9 provide written examples and stroke guides; the advanced parts provide dictionary readings and selected opening examples without audio. Japanese characters use self-hosted subsets of Noto Sans JP, licensed under the SIL Open Font License. Stroke data loads only when a guide is opened. Five rare forms use KanjiVG-derived glyph images when the font lacks them. The form 屮 has no verified animation in this release; its card says so and does not offer a broken stroke button.

Practice, not certification

Test scores reflect the answers in the current tab. Twelve consecutive correct answers count as session mastery; a missed target returns its streak to zero. Directions have separate scores. Refreshing the page clears progress. Writing practice has no handwriting recognition or grading.

Suggest a correction if a reading, example or character needs attention.