Dataset note
Word and adjective data: scope, sources, and limits
All 18 published Word and Adjective corpora are exact-locale releases. The table separates release-review status from native-editor review.
Current status
All 18 published corpora passed the defined release review. This does not mean a native-editor review. Exact-locale requests never fall back to English.
Sources and license
European, English, Chinese and Korean releases use lemma and part-of-speech annotations from pinned Universal Dependencies snapshots. Japanese uses a pinned JMdict_e snapshot. Both source families are attributed under CC BY-SA 4.0; RandomHub publishes filtered, modified extracts.
| Locale | Words | Adjectives | Version | Status | Source | License |
|---|---|---|---|---|---|---|
| en | 700 | 105 | 2026-07-24-r1 | Words: Reviewed Adjectives: Reviewed | UniversalDependencies/UD_English-EWT | CC BY-SA 4.0 |
| fr | 1,329 | 148 | 2026-07-24-r1 | Words: Reviewed Adjectives: Reviewed | UniversalDependencies/UD_French-GSD | CC BY-SA 4.0 |
| de | 841 | 142 | 2026-07-24-r1 | Words: Reviewed Adjectives: Reviewed | UniversalDependencies/UD_German-GSD | CC BY-SA 4.0 |
| ru | 399 | 114 | 2026-07-24-r1 | Words: Reviewed Adjectives: Reviewed | UniversalDependencies/UD_Russian-GSD | CC BY-SA 4.0 |
| ja | 287 | 74 | 2026-07-24-r2 | Words: Reviewed Adjectives: Reviewed | Electronic Dictionary Research and Development Group (EDRDG), JMdict_e | CC BY-SA 4.0 |
| zh-hant | 280 | 70 | 2026-07-24-r2 | Words: Reviewed Adjectives: Reviewed | UniversalDependencies/UD_Chinese-GSD | CC BY-SA 4.0 |
| ko | 275 | 79 | 2026-07-24-r2 | Words: Reviewed Adjectives: Reviewed | UniversalDependencies/UD_Korean-GSD | CC BY-SA 4.0 |
| es | 1,380 | 172 | 2026-07-24-r1 | Words: Reviewed Adjectives: Reviewed | UniversalDependencies/UD_Spanish-GSD | CC BY-SA 4.0 |
| pt | 291 | 79 | 2026-07-24-r2 | Words: Reviewed Adjectives: Reviewed | UniversalDependencies/UD_Portuguese-Bosque | CC BY-SA 4.0 |
Language-specific filters
Latin and Cyrillic tools use written prefix/suffix and grapheme length. Japanese also supports reading, mora count, and i/na adjective classes. Traditional Chinese pinyin is disabled because the current readings are not reliable. Korean uses Hangul syllable count.