Kanji Stroke Order Animator
Watch any kanji or kana drawn stroke by stroke in the Japanese order, with stroke numbers, stroke types and components.
Kanji Stroke Order Animator
Stroke data: KanjiVG © Ulrich Apel, CC BY-SA 3.0 — the data files served here are adapted from it under the same licence.
Stroke-order diagram
One frame per stroke — the new stroke in colour.Works in your browser; each character's stroke data is a small file served from this site. Japanese stroke order — Chinese teaching differs for some characters, which is why this is a separate tool from the Chinese stroke order animator.
How to use the kanji stroke order animator
Enter one character, or up to eight — a word or a name. Each gets a tab. Half-width katakana are converted to full width automatically. The data covers 6,702 characters: the kanji of Japan's JIS level 1 and level 2 character sets, plus hiragana and katakana.
Play draws the strokes in order at the speed you choose, the current stroke in colour. Next stroke steps through one at a time, which is the better way to learn a complicated character. Numbers mark where each stroke begins.
The panel beside the canvas lists the character's components — 字 is 宀 over 子, and 子 is the phonetic — and every stroke with its Unicode stroke type. Click a component to see which strokes belong to it.
Below the canvas, one small frame per stroke shows the character building up. It prints cleanly for a practice sheet.
Japanese stroke order: where it comes from, and where it differs from Chinese
Stroke order is the sequence in which the strokes of a character are written, and schools teach one order per character so that learners write consistently. Japanese schools teach an order based on the Ministry of Education's 1958 guide, 筆順指導の手びき, drawn up so that teachers of the then 881 education kanji would not contradict one another. Its rules are the familiar ones: top to bottom, left to right, a horizontal before a vertical that crosses it — with named exceptions.
右 and 左: the rule that trips learners
The best-known exception pair is 右 and 左. The guide's principle is that where the horizontal stroke is long and the left-falling stroke short, the left-falling stroke comes first; where the horizontal is short and the left-falling stroke long, the horizontal comes first. So 右 begins with ノ and 左 begins with 一 — and that is exactly how the data behind this page draws them. Stroke order is also not identical across countries: China and Taiwan have their own standards, and 右 is one of the characters Japan and China teach differently. That is why this is a separate tool from our Chinese stroke order animator, not a setting on it.
The 1958 guide says of its own orders that they were fixed to avoid confusion in teaching, and that this does not make other orders wrong. — Ministry of Education, 筆順指導の手びき (1958), quoted in 「筆順」
More than one right answer
The guide is candid about this. It lists characters that have two widely used orders — among them 上, 必, 取, 発, 馬 and 無 — and says that choosing one for teaching does not make the other a mistake. This page shows the order in KanjiVG, an open dataset whose shapes follow Japanese schoolbook fonts and whose stroke order "follow[s] the Japanese example in all cases". Where you have learned a different accepted order, both may be correct.
What the strokes are made of
Every stroke on this page carries a type from Unicode's CJK Strokes block — ㇐ for a horizontal line, ㇑ for a vertical, ㇒ for a stroke falling to the left, ㇔ for a dot, and compound types for strokes that turn or end in a hook. Unicode names them with letters taken from their Chinese names: a plain horizontal is CJK STROKE H, the turning hook on the right of 印 is CJK STROKE HZG. Across all 79,907 strokes in the data, the plain horizontal is the most common, more than one stroke in four. Characters are also built from components, and the data marks which is the radical and which carries the sound: in 字, the roof 宀 sits over 子, and 子 is the phonetic.
Learning kanji in Southeast Asia
Japanese is widely studied across Southeast Asia, and many learners there already read Chinese characters. For them the traps are not unfamiliar characters but familiar ones written in a different order or shape — which is exactly what a Japanese-order animator is for.
Why 右 and 左 start differently, a 1958 guide that allows two answers, and the 30-stroke 驫
Japanese school stroke order comes from the Ministry of Education's 1958 guide 筆順指導の手びき, written for the 881 education kanji of the time.
The guide itself says the orders it chose are for consistency in teaching and do not make other orders wrong.
右 begins with the left-falling stroke and 左 with the horizontal: where the horizontal is long and the left-falling stroke short, the left-falling stroke goes first.
The guide names characters with two widely used orders, including 上, 必, 取, 発, 馬 and 無.
Japan and China teach different orders for some characters; 右 is one of them.
The data behind this page, KanjiVG, covers 6,702 characters, including hiragana and katakana, drawn from Japanese schoolbook fonts.
Those 6,702 characters contain 79,907 strokes between them.
驫, three 馬 stacked together, has 30 strokes — the most in the data.
The plain horizontal stroke ㇐ is the most common type: more than one stroke in four across the whole set.
Unicode names each stroke type with letters from its Chinese name, so the turning hook on the right of 印 is CJK STROKE HZG.
Frequently asked questions
-
Mostly, but not always. Japan, mainland China and Taiwan each have their own school standards, and some characters — 右 is a well-known one — are taught differently. This page shows the Japanese order; our Chinese stroke order animator shows the Chinese one.
-
From KanjiVG, an open dataset by Ulrich Apel whose glyphs are based on Japanese schoolbook fonts and whose stroke order follows the Japanese example in all cases. Japanese school practice itself goes back to the Ministry of Education's 1958 guide 筆順指導の手びき.
-
Because of a rule in the 1958 guide: where the horizontal stroke is long and the left-falling stroke short (右), the left-falling stroke comes first; where the horizontal is short and the left-falling stroke long (左), the horizontal comes first.
-
Not necessarily. The 1958 guide says its orders do not make other established orders wrong, and it lists characters with two widely used orders, such as 必, 上 and 馬.
-
6,702 characters, which KanjiVG describes as the kanji of the Japanese JIS level 1 and level 2 standards, plus hiragana, katakana and a few symbols. If a character is not in the data, the page says so rather than guessing.
-
They are characters from Unicode's CJK Strokes block, which KanjiVG uses to record each stroke's shape — ㇐ a horizontal, ㇑ a vertical, ㇒ a left-falling stroke, ㇔ a dot, and compound types for strokes that turn or hook. The page shows each type's Unicode name, such as CJK STROKE HZG.
-
Most kanji are built from smaller parts. The panel lists them with their positions, marks which part dictionaries use as the radical, and which part suggests the pronunciation. Different dictionaries sometimes choose different radicals, so the page says which tradition each one follows.
-
Yes. The stroke-order diagram below the canvas shows one frame per stroke and is designed to print cleanly for a practice sheet.
-
The data comes from KanjiVG under the Creative Commons Attribution-ShareAlike 3.0 licence, and the files this page serves are adapted from it under the same licence. Credit KanjiVG and Ulrich Apel with a link to kanjivg.tagaini.net.
-
No. It fetches the stroke data for the character you type from this site; what you type is not sent anywhere else or stored.
Method & sources
What this tool implements
- Stroke order, shapes and components are KanjiVG's, which follows "Japanese schoolbook fonts" and "the Japanese example in all cases". The data is generated, never hand-edited: the build pins the release by SHA-256 and fails on a missing, out-of-order or unnumbered stroke.
- ⚠️ One upstream defect is repaired and recorded rather than hidden: KanjiVG r20250816 gives 參 only 11 stroke-number labels for its 13 strokes; the two missing labels are placed at their strokes' starting points and the character is listed in index.json. Strokes typed with a stray character upstream (６ on 11 strokes, 一 on one) show no type rather than a guessed one.
- The test suite runs against all 6,702 generated files (every stroke drawable and labelled, every component range valid, totals matching the index) and pins the Japanese order the 1958 guide describes for 右 (ノ first) and 左 (一 first).
- Where the 1958 Ministry of Education guide accepts more than one order (上, 必, 馬 …), the page says so; it shows KanjiVG's order without claiming it is the only correct one.
- Licence: KanjiVG is CC BY-SA 3.0. The derived files carry the credit in every file, a LICENSE.txt sits beside them, and the page credits KanjiVG and Ulrich Apel with links under the canvas.
Sources
- KanjiVG r20250816, Ulrich Apel, CC BY-SA 3.0: https://github.com/KanjiVG/kanjivg/releases/tag/r20250816 — project and file documentation: https://kanjivg.tagaini.net/files.html
- KanjiVG, Stroke types (the kvg:type values and their Unicode CJK Strokes code points): https://kanjivg.tagaini.net/stroke-types.html
- Unicode Character Database 18.0, UnicodeData.txt (CJK STROKE names, U+31C0–U+31E5): https://www.unicode.org/Public/UCD/latest/ucd/UnicodeData.txt
- 文部省『筆順指導の手びき』(1958), its principles and its statement that its orders do not make other orders wrong, as quoted in the Japanese Wikipedia article 「筆順」 (citing 沖森卓也, 2011): https://ja.wikipedia.org/wiki/%E7%AD%86%E9%A0%86
What can make this go out of date
- None at runtime beyond this site: each character's stroke data is a static file under /tools/_data/kanjivg/, generated from KanjiVG and committed to the repository. Nothing the reader types is transmitted.
Pick up where you left off
Stored only in this browser — never sent to our servers.