URL slugs — three ways to turn a title into an address
A slug is the readable part at the end of a URL: example.com/blog/here. For an English title the job is small — lowercase it and swap spaces for hyphens. For a title in any other script there is a decision to make first, and it does not have one right answer, so this tool shows all three.
A slug does two things. It lets a reader guess what a page is about from the address alone, and it gives a search engine one more signal about the topic. That is why /blog/seoul-restaurant-guide beats /blog/1382.
The complication is that a URL may only contain a small set of characters, and letters outside the Latin alphabet are not among them. A slug that keeps its own script is percent-encoded on the way to the server, so an eight-character slug can become a fifty-six-character address. Delete those characters instead and there may be no slug left at all. Choosing which cost to pay is what the three buttons are for.
One title, three results
The Korean title 2026년 블로그 SEO 가이드 through each mode. The last column is the length after percent-encoding.
| Mode | Slug | Length in a URL |
|---|---|---|
| Keep script | 2026년-블로그-seo-가이드 | 73 |
| Romanize | 2026nyeon-beullogeu-seo-gaideu | 30 |
| Latin only | 2026-seo | 8, and the meaning is gone |
Is a non-Latin slug safe to use?
Yes. Google indexes them normally and displays them decoded in search results, and major Korean and Japanese blogging platforms use them by default.
There are three costs. Copying the link into a chat window or a document sometimes exposes the encoded form. Every character takes three bytes, so nine characters of URL, which matters where a length limit applies. And older forum software occasionally mishandles them.
For a readership that arrives through search and social links, keeping the original script is comfortable. For documentation, where addresses get typed by hand or pasted into code, romanized slugs are easier to live with.
Romanization transcribes sound, not letters
Most tools map one letter to one letter. That turns 종로 into “jongro”, but the correct transcription is Jongno. The Revised Romanization of Korean is defined in terms of pronunciation, and Korean pronunciation changes where two syllables meet — the ㄹ becomes ㄴ after the preceding ㅇ.
This tool implements the rules from the official notice: liaison, nasalization, lateralization, palatalization, and aspiration with ㅎ. Tensification is left out because the notice explicitly says not to reflect it, which is why 압구정 is Apgujeong rather than Apkkujeong.
Where letter-by-letter mapping goes wrong
The middle column is what a substitution table produces; the right column is the spelling printed in the official notice.
| Korean | Letter by letter | Correct |
|---|---|---|
| 종로 | jongro | Jongno |
| 신라 | sinra | Silla |
| 대관령 | daegwanryeong | Daegwallyeong |
| 왕십리 | wangsipri | Wangsimni |
| 백암 | baekam | Baegam |
| 울릉 | ulreung | Ulleung |
| 해돋이 | haedodi | haedoji |
| 좋고 | johgo | joko |
Some cases the rules cannot settle alone
The notice itself assumes knowledge this tool does not have. When ㄴ meets ㄹ, the result is ll inside a single word (신라 Silla) and nn across a compound boundary (신문로 Sinmunno). Telling those apart requires knowing how the word was formed, which is a dictionary's job.
The tool applies the Silla rule. If you are turning a place name or a proper noun into a slug, check the result. The same applies to nouns such as 묵호 (Mukho) where the ㅎ is written out rather than merged.
The same text can be stored two different ways
Text typed on a Mac often arrives decomposed: an accented letter, or a Korean syllable, is stored as several code points rather than one. It looks identical on screen and compares as different.
Build a slug from that directly and the same title produces two different addresses depending on which machine it was copied from — and a decomposed Korean syllable is three times longer once encoded. This tool normalizes the input first, so the result does not depend on where the text came from.
What makes a slug worth having
- •Three or four meaningful words rather than the whole title. Long slugs get truncated in search results.
- •Hyphens, not underscores — the long-standing advice is that search engines treat a hyphen as a word boundary and an underscore as part of the word.
- •No year or date if you plan to update the page. Editing a 2025 URL in 2026 breaks every link to it.
- •Do not change a slug once it is published. If you must, redirect the old address to the new one with a 301.
- •Lowercase only. Some servers treat differently-cased addresses as different pages.
Frequently asked questions
Do non-Latin slugs hurt rankings?
No. Google has said encoded URLs are handled normally. The real difference is not ranking but how the link looks and how long it gets when copied.
The romanized spelling is not the one I recognize.
Personal names conventionally keep whatever spelling their owner already uses (이 as Lee, 김 as Kim) rather than following the notice. Some place names have established spellings too, so check proper nouns by hand.
What happens to emoji?
They are not letters, so they become separators. Emoji in a URL is technically possible but tends to break when links are shared, so it is not worth the trouble.
How long should a slug be?
There is no hard limit, but staying under about 60 characters keeps it from being cut off in search results.