Generate printable flash cards, handwriting practice sheets, reference charts, and stroke order guides for all hiragana and katakana characters.
Produces five types of PDF:
- Flash cards — 208 double-sided cards (104 hiragana + 104 katakana) arranged in a 3×3 grid on A4 pages, designed for duplex printing and cutting.
- Practice sheets — Landscape A4 pages with 2cm grid boxes. Each character gets 3 rows: the first box shows the KanjiVG stroke-order character, the rest are empty for handwriting practice. Boxes include dashed cross guides for centering.
- Reference charts — Landscape A4 pages with the standard gojūon table layout: basic characters, dakuten/handakuten, and yōon combinations with romaji labels.
- Stroke order — Portrait A4 pages with KanjiVG stroke diagrams and numbered indicators for each character, split into basic, dakuten, and combination sections for readability.
- Vocabulary reference — Per-JLPT-level (N5–N1) vocabulary PDFs in English and Swedish, sourced from JMdict. Each level contains all words for that level with meanings, POS, register tags, and antonyms.
pip install -r requirements.txt
python generate.py --setup # verify prerequisites
python generate.py --kana # generate flash cards → output/kana_flashcards.pdfRequires Python 3.10+ and ReportLab. The Klee One font is bundled.
Download the latest pre-built PDFs:
- Flash cards 55×82mm
- Flash cards 63×88mm
- Flash cards 74×105mm
- Practice sheets — Hiragana
- Practice sheets — Katakana
- Reference chart
- Reference chart — Hiragana
- Reference chart — Katakana
- Stroke order
- Stroke order — Hiragana
- Stroke order — Katakana
- Vocabulary N5 (EN)
- Vocabulary N5 (SV)
- Vocabulary N4 (EN)
- Vocabulary N4 (SV)
- Vocabulary N3 (EN)
- Vocabulary N3 (SV)
- Vocabulary N2 (EN)
- Vocabulary N2 (SV)
- Vocabulary N1 (EN)
- Vocabulary N1 (SV)
Flash cards — front (character) and back (romaji + stroke order):
Practice sheet:
Reference chart:
Stroke order:
Vocabulary reference:
Each script includes:
- 46 basic gojūon (あ→ん / ア→ン)
- 20 dakuten variants (が, ざ, だ, ば / ガ, ザ, ダ, バ)
- 5 handakuten variants (ぱ行 / パ行)
- 33 yōon combinations (きゃ, しゅ, ちょ, etc.)
| Front | Back |
|---|---|
| Large kana character | Romaji reading |
| Type label (hiragana/katakana) | Stroke order diagram |
Cards are 55×82mm (roughly A7), printed 9 per A4 sheet with cut lines.
All commands go through python generate.py. Run without arguments to see help.
python generate.py --kana # Flash cards (208 cards)
python generate.py --kana --hiragana-only # Hiragana only
python generate.py --kana --katakana-only # Katakana only
python generate.py --kana -o my_cards.pdf # Custom output path
python generate.py --kana --practice # Practice sheets
python generate.py --kana --practice --hiragana-only
python generate.py --kana --chart # Reference chart
python generate.py --kana --chart --hiragana-only
python generate.py --kana --stroke-order # Stroke order guide
python generate.py --kana --stroke-order --hiragana-only
python generate.py --kana --card-width 63 --card-height 88 # Custom card size (mm)python generate.py --vocabulary --jlpt n5 # N5 English → output/vocabulary_n5.pdf
python generate.py --vocabulary --jlpt n5 --lang sv # N5 Swedish → output/vocabulary_sv_n5.pdf
python generate.py --vocabulary --jlpt n3 # N3 English → output/vocabulary_n3.pdf
python generate.py --anki --jlpt n5 # N5 Anki deck → output/anki_n5.apkg
python generate.py --anki # All levels → output/anki_n5.apkg … anki_n1.apkg + anki_all.apkgpython generate.py --setup # Check prerequisites; auto-build vocabulary data if missing
python generate.py --build-vocabulary # Force rebuild data/vocabulary.json (~50MB download)python generate.py --update-strokes # Re-download KanjiVG and regenerate stroke data
python generate.py --generate-examples # Regenerate example images in docs/ (requires poppler-utils)- Print the PDF double-sided with long-edge flip
- Cut along the gray grid lines
- Each card's front (character) aligns with its back (romaji + strokes)
Adjust BACK_PAGE_OFFSET_Y in config.py if your printer's duplex alignment is off.
All tunables are in config.py. Values use ReportLab's mm unit.
| Variable | Default | Description |
|---|---|---|
CARD_WIDTH |
55mm | Card width (also settable via --card-width) |
CARD_HEIGHT |
82mm | Card height (also settable via --card-height) |
BACK_PAGE_OFFSET_Y |
-2.0mm | Vertical shift for back page content to compensate for duplex misalignment. Positive = up, negative = down |
KANA_FONT_SIZE |
34mm | Large character size on card front |
ROMAJI_FONT_SIZE |
8mm | Romaji text size on card back |
LABEL_FONT_SIZE |
3mm | "hiragana"/"katakana" label size on card front |
STROKE_LINE_WIDTH |
0.8mm | Stroke diagram line thickness |
STROKE_DOT_RADIUS |
1.0mm | Stroke start-point dot size |
STROKE_COLORS |
list of 6 RGB tuples | Per-stroke colors cycling indigo→emerald→red→amber→violet→cyan |
CUT_LINE_WIDTH |
0.2mm | Cut guide line thickness |
CUT_LINE_COLOR |
(0.7, 0.7, 0.7) | Cut guide line color (light gray) |
Grid layout (COLS, ROWS, MARGIN_X, MARGIN_Y) is auto-computed from the card and page sizes so cards are always centered on A4.
Stroke order data derived from KanjiVG by Ulrich Apel, licensed under Creative Commons Attribution-ShareAlike 3.0.
Adaptations: SVG stroke paths extracted, dakuten/handakuten/yōon variants composed from base characters for flash card and practice sheet rendering.
data/vocabulary.json is built from two sources:
JLPT word lists by Jonathan Waller (https://www.tanos.co.uk/jlpt/), used with permission. Lists cover N5–N1 vocabulary.
JMdict by the Electronic Dictionary Research and Development Group (https://www.edrdg.org/), licensed under Creative Commons Attribution-ShareAlike 3.0. Provides multilingual meanings, part of speech, register tags, and antonyms.
- Klee One SemiBold by Fontworks — used for all Japanese text and labels. Licensed under the SIL Open Font License 1.1.
| Package | License | Use |
|---|---|---|
| ReportLab | BSD | PDF generation |
| pykakasi | GPL-3.0-or-later | Romaji generation (build-time only) |
| genanki | MIT | Anki deck export |