CJK Name & Address Romanizer
Korean Hangul → Revised Romanization (the South Korean government standard), with a road-name address mode for international shipping. Japanese kana → MOFA passport Hepburn (外務省ヘボン式). Both implement the actual published rule sets, including the parts that a per-syllable lookup table gets wrong.
Korean: why 신라 is Silla and not “Sinra”
The Revised Romanization standard says something most converters quietly ignore. Chapter 3, Article 1 requires you to write the result of the sound change, not the letters as spelled. A naive syllable-by-syllable lookup gives “Sinra”; the standard gives Silla, because ㄴ followed by ㄹ is pronounced ㄹㄹ. This tool decomposes each syllable into its jamo and applies the real phonological rules before romanizing.
| Rule | Hangul | Naive lookup | Correct (RR) |
|---|---|---|---|
| 비음화 nasalization | 백마 | Baekma | Baengma |
| ㄹ → ㄴ after a stop | 종로 | Jongro | Jongno |
| stacked assimilation | 왕십리 | Wangsipri | Wangsimni |
| 유음화 lateralization | 신라 | Sinra | Silla |
| 유음화 (reverse) | 별내 | Byeolnae | Byeollae |
| compound cluster | 독립문 | Dokripmun | Dongnimmun |
| 구개음화 palatalization | 해돋이 | Haedodi | Haedoji |
| 구개음화 | 같이 | Gati | Gachi |
| ㅎ aspiration | 좋고 | Jotgo | Joko |
| ㅎ kept in nouns | 묵호 | Muko | Mukho |
| 연음 resyllabification | 백암 | Baekam | Baegam |
| ㄴ insertion | 학여울 | Hakyeoul | Hangnyeoul |
Every one of those pairs is a worked example printed in the standard itself, and each is used as a regression test on the implementation. Two of the rules are genuinely lexical rather than derivable — the ㄴ-insertion in 학여울 and the handful of Sino-Korean compounds like 신문로 that take ㄴㄴ instead of ㄹㄹ — so those are handled from an explicit list rather than guessed at.
Japanese: the three rules that define MOFA Hepburn
Passport Hepburn is not the same as the Hepburn you learned from a textbook. The Ministry of Foreign Affairs specifies three transformations, and they are where nearly every mistake on a passport application happens:
N normally, but M before B, M and P.
難波 NAMBA · 本間 HOMMA · 三瓶 SAMPEI
Doubles the next consonant — except before CH, where it becomes T.
服部 HATTORI · 吉川 KIKKAWA · 八丁 HATCHO
Long O and U are simply not written. No extra O, no extra U, no macron.
大野 ONO · 河野 KONO · 庄司 SHOJI · 裕子 YUKO
Since 1 April 2000 MOFA also permits writing a long O as OH — 大野 may be ONO or OHNO. The catch is that once you choose one, you must use it consistently on every future application, so the tool shows both and labels the alternative rather than picking for you.
Preguntas frecuentes
영문주소 변환은 EMS 국제우편에 그대로 쓸 수 있나요?
네. 주소 모드는 행정안전부의 도로명주소 영문 표기 규칙을 그대로 따릅니다. 도로 위계에 따라 붙임표를 넣어 -daero(대로), -ro(로), -gil(길)로 적고, 행정구역은 -si, -gun, -gu, -eup, -myeon, -ri, -dong, -ga로 적으며, 영문은 작은 단위부터 큰 단위 순으로 뒤집어 씁니다. 예를 들어 「서울특별시 종로구 세종대로 110」은 「110, Sejong-daero, Jongno-gu, Seoul, Republic of Korea」가 됩니다. 우편번호 5자리를 함께 입력하면 국가명 바로 앞에 넣어 줍니다. 다만 건물번호와 동·호수는 배송에 가장 중요한 부분이므로 원문과 반드시 대조해 주세요.
ヘボン式 ローマ字 変換 はパスポート申請にそのまま使えますか？
外務省が定めるヘボン式の規則をそのまま実装しています。撥音「ん」はB・M・Pの前でMになり（NAMBA・HOMMA・SAMPEI）、促音「っ」は次の子音を重ねますが「ち・ちゃ・ちゅ・ちょ」の前だけTになり（HATTORI・KIKKAWA・HOTCHI・HATCHO）、長音の「オウ」「オオ」「ウウ」はOやUを入れずに書きます（ONO・KONO・SHOJI・YUKO）。2000年4月1日以降認められているOH表記（大野→OHNO）も切り替えられます。ただし一度どちらかを選んだら以後の申請では変更できないため、実際の提出前に必ず旅券窓口の表記案内でご確認ください。
Why does the tool ask for kana instead of accepting kanji?
Because kanji do not have a single reading. 大野 can be Ono or Ohno, 東 can be Higashi or Azuma or Higashi-, and 幸子 can be Sachiko or Yukiko — the reading is a fact about your family, not about the characters. A converter that guessed would produce a wrong passport name, which is exactly the failure this tool exists to avoid. Passport applications themselves ask for the フリガナ reading for the same reason. If you paste kanji, the tool flags each character rather than inventing a reading for it.
Why does my Korean surname come out as Gim, I or Bak?
Those are the strictly correct Revised Romanization forms of 김, 이 and 박 — but almost nobody uses them. Chapter 3, Article 7 of the standard explicitly allows an established personal-name spelling to stand, which is why Korean passports say Kim, Lee and Park. The tool shows both: the standard form, and the conventional passport spellings actually in use. For a passport, a visa or an airline ticket, use whatever is already printed on your existing documents and keep every document consistent.
What is the hyphen option in Korean name mode for?
Chapter 3, Article 2 permits an optional hyphen where a romanized syllable boundary would otherwise be ambiguous. Without it, 중앙 becomes 'Jungang', which could be read as jung-ang or jun-gang; with it, you get the standard's own example Jung-ang. Same for 반구대 → Ban-gudae, 세운 → Se-un and 해운대 → Hae-undae. It is off by default because most Korean addresses and signs are written without it, but turn it on when the reading genuinely matters. Note the standard treats it as optional, so both forms are correct.
Does anything get uploaded to a server?
No. Both romanizers are pure JavaScript running in your browser — the Hangul jamo decomposition, all the Korean phonological rules, and the whole MOFA Hepburn kana table. There is no AI model and no network request, which is the right design here: these are published, deterministic government algorithms, so a lookup plus rule engine is both more accurate and more private than asking a model to guess.