Licenses & attribution

Our dictionary and grammar data comes from open sources with real conditions attached. We keep those sources in separate zones so that each obligation applies where it should, and we publish the part that carries a share-alike obligation.

Zone A — Korean dictionary data (CC BY-SA 2.0 KR)

Headwords, senses, English glosses, pronunciations, conjugated forms and speech-level tags are derived from:

• Work: Korean Basic Dictionary (국립국어원 한국어기초사전, krdict), National Institute of Korean Language — krdict.korean.go.kr
• Licence: Creative Commons Attribution-ShareAlike 2.0 Korea (CC BY-SA 2.0 KR)
• Changes we made: we extracted a subset of fields, normalised them and rebuilt them as a database. We did not use the audio recordings, which carry separate terms.

The share-alike download

ShareAlike means the derived dataset must be offered under the same licence. It is, as a single archive containing the Zone A tables only, with the attribution and the list of changes inside it.

Not published yet. The download endpoint goes live with the service — the dataset is built and served by our API server, which is not deployed yet. Until then nothing has been distributed, so no share-alike obligation has been triggered. Ask us at our contact address if you need it sooner.

What else Zone A covers

The pronunciation exception dictionary bundled inside the extension is also derived from krdict, and is covered by the same publication above.

Zone B — Grammar explanations (CC BY 4.0)

• Work: kimchi-grammar
• Licence: Creative Commons Attribution 4.0 International (CC BY 4.0)
• Attribution only — no share-alike. We keep it physically separate from Zone A on purpose: merging the two would pull the result under ShareAlike. They are joined at display time, never in storage.
• Corrections we make to these entries are offered back upstream.

Zone C — Our own work

The romanization and pronunciation rule engines, the speech-level rules, the overlay, the extension and this website are ours. The Korean standard pronunciation and romanization rules they implement are public rules, not someone's dataset.

Software

• Morphological analysis uses Kiwi (LGPL v3), run as a separate service rather than linked into our code.

Subtitles

Subtitles belong to whoever produced them. Your browser fetches them from the platform you are watching; we analyse them and show the result back to you. We do not store them and we do not redistribute them. See the privacy policy for the detail.