Independent dictionary layers
Additional readings and source senses, kept distinct from CC-CEDICT, HanDeDict and CC-Canto.
Rime Cantonese
Cantonese Computational Linguistics Infrastructure Development Workgroup (CanCLID) and Rime Cantonese contributors. Original project · CC-BY-4.0.
Pinned commit: 259f0e48bba840c3a2e0d117539e96937f3d89bc. 128,139 written forms. Cantonese reading lexicon; not definitions or usage frequencies.
Exact character and word spellings with numbered Jyutping selected from two pinned files; duplicates collapsed with source lines retained. Input weights excluded. ODbL maps, phrases and lettered files not imported.
Full license · Attribution & reuse · Edition & checksums · Download source data pack
English Wiktionary · Chinese definitions
Wiktionary contributors; per-entry revision and history links identify the source and contributors. Original project · CC-BY-SA-4.0.
173,847 written forms. Chinese-script spellings with English definitions from the raw English Wiktionary extraction, supplemented by revision-pinned API records. Coverage depends on source availability and import filters.
Selected Chinese part-of-speech sections converted to plain-text definitions; usage labels retained, nested definitions retain parent context. Examples, quotations, media, references and verification-flagged senses excluded. Source section hashes and revision IDs retained for API records; raw extraction line numbers and record hashes retained for bulk records. Bulk usage tags, topics and etymology grouping retained. The curated API records take precedence for their spellings. No inferred reading or variety alignment. No automated translation.
Bulk extraction by Wiktextract / Kaikki.org, from the 2026-09-02 Wiktionary dump. Snapshot attribution is used where no page revision ID is supplied.
Full license · Attribution & reuse · Edition & checksums · Download source data pack
These source layers keep their own licenses. zimai’s application copyright does not restrict reuse of this licensed content. Wiktionary-derived content is available under CC BY-SA 4.0. No endorsement by either source is implied.