URL Slug Generator

What this tool does

A slug is the human-readable part of a URL that identifies a page: the url-slug-generator in /url-slug-generator/. Getting it right matters more than it looks. A slug with accents, spaces or punctuation has to be percent-encoded in the address bar, which makes it ugly and can break when the link is pasted into chat apps, email clients or older systems. A slug that keeps its accents in one place and loses them in another produces two URLs for the same page, which splits search signals.

Method

The phrase is decomposed with Unicode NFD so that accented letters separate into a base letter plus a combining mark; the marks are then removed, which folds é to e and ệ to e without a lookup table. A small additional map handles the letters that do not decompose and have a conventional Latin equivalent — đ→d, ß→ss, ø→o, æ→ae, œ→oe, ł→l, þ→th, ı→i. The result is lower-cased, apostrophes are dropped so don’t becomes dont, every remaining run of non-alphanumeric characters becomes a single hyphen, and leading and trailing hyphens are trimmed.

Example

Input	Slug
10 mẹo hay cho Đà Nẵng — 2026	10-meo-hay-cho-da-nang-2026
Café & Crème Brûlée	cafe-creme-brulee
Don’t Panic! (A Guide)	dont-panic-a-guide
Straße Übersicht	strasse-ubersicht

What a slug has to survive

A slug is the human-readable part of a URL. It is not just “the title with spaces replaced” — it has to survive being copied out of a browser, pasted into a chat application, put in an email, and read aloud. That imposes a stricter character set than a filename:

• Lowercase ASCII letters, digits and hyphens. Nothing else is safe without escaping, and escaped characters in a path are unreadable.
• Hyphens, not underscores. Google treats hyphens as word separators; underscores join words into one token.
• No leading or trailing hyphen, no doubled hyphens. They look like a mistake and some routers normalise them inconsistently.
• No percent-encoding. A slug containing %20 is a slug that was not finished.

Accents: transliterate, do not strip

There are two ways to remove an accent, and only one of them is right. Stripping the diacritic turns café into caf, which loses a letter. Transliterating turns it into cafe, which is the ASCII spelling a reader expects. This tool decomposes with NFD, removes the combining marks, and keeps the base letter — so Café becomes cafe, not caf.

German ß is the special case: it has no decomposed form, so it has to be mapped explicitly to ss, which is what German orthography does when the character is unavailable. Straße → strasse, not strae.

Vietnamese is the hardest Latin-script case because one letter can carry two diacritics — Đà Nẵng has a base change (Đ) and tone marks on two vowels. The base change has to be mapped (Đ → d) and the tone marks removed, giving da-nang. The same logic is available as a dedicated tool if you need it without the slug formatting.

The characters people expect to survive

Character	Slug result	Reason
apostrophe (’ or ')	removed	Don't → dont, the common convention
ampersand	removed or “and”	This tool removes it: Café & Crème → cafe-creme
plus	hyphen	It means a space in a query string, so it becomes a word break: a+b → a-b
percentage	removed	It starts a percent-escape
slash	hyphen	A literal slash would create a new path segment: a/b → a-b
digits	kept	Years and ids are meaningful and safe

Length and uniqueness

Slugs are not required to be unique by any standard, but two different titles can produce the same slug — Café & Crème and Cafe Creme both become cafe-creme. If the slug is a key, you need a collision strategy: append a short id or a counter, and check before publishing. If the slug is only cosmetic (the id is the key), a collision is harmless.

Long titles produce long slugs. There is no hard limit, but practical limits exist: some systems cap path segments, and a slug beyond about 60–75 characters is unlikely to be read in full. Truncating at a word boundary and dropping a trailing hyphen — the failure mode of a naive cut — is safer than truncating mid-word.

Non-Latin scripts

A title written entirely in Chinese, Japanese, Cyrillic or Arabic has no Latin base letters to keep, so transliteration-by-decomposition produces an empty string. This tool reports an empty result rather than inventing a transliteration, because choosing one (pinyin? Hepburn? ISO 9?) is a language decision, not a formatting one.

For those titles the practical options are: keep the original characters in the URL (modern browsers display them, and the server percent-encodes them), or supply a manual Latin slug. Note that an empty slug concatenated into a path produces a double slash — worth checking for before you publish.

Slug versus title, in search

The slug is a ranking signal, but a weak one. Words in the slug are treated like words in the URL, and they help a reader decide whether to click. That argues for a readable slug over a keyword-stuffed one: how-to-convert-base64-to-text reads better and ranks no worse than base64-converter-online-free-tool.

Frequently asked questions

Go deeper

Longer explanations of the problems this tool solves:

Cleaning text that looks right and behaves wrong

Sources

• Unicode Standard, normalization form NFD and the combining diacritical marks U+0300–U+036F.
• RFC 3986, Uniform Resource Identifier (URI): Generic Syntax, for the characters that are safe in a URL path.