Slug Generator

A URL slug is the human-readable identifier in a URL — the `introduction-to-typescript` portion of `example.com/blog/introduction-to-typescript`. Slugs serve three audiences at once. Search engines use them as a structural hint about page content. Humans use them when deciding whether to click a shared link or trust a URL pasted into a chat. And maintainers use them as stable references that survive title edits. A good slug is short, lowercase, ASCII-only, and unambiguous; a bad slug is a randomly-generated UUID or the literal article title with spaces left in.

The hard part of slug generation is what to do with characters that are not ASCII. A naïve "lowercase + replace non-alphanumeric with hyphen" approach loses every accent, every non-Latin character, and most punctuation that carries meaning (`&`, `@`). The slugifier here does the standard three-step dance used by libraries like `slugify`, `@sindresorhus/slugify`, and Django's `slugify` utility: first Unicode NFKD-normalize the string, which splits accented letters into a base letter plus a combining mark; second, drop the combining marks (turning `é` into `e`, `ñ` into `n`); third, apply hand-coded transliterations for letters that do not decompose to ASCII (`ß → ss`, `æ → ae`, `ø → o`, `ł → l`, `þ → th`). Symbols with semantic meaning get spelled out: `&` becomes `and`, `@` becomes `at`.

Three options matter in practice. **Separator** — hyphens are the convention for URLs because search engines read them as word boundaries, while underscores are not. Filenames sometimes prefer underscores for shell-friendliness. **Stop-word stripping** is useful for category and tag slugs ("the-best-cameras" reads better as "best-cameras") but should usually stay off for article slugs where the full title needs to be recognizable. **Max length** keeps long titles from producing 200-character slugs that get truncated in search result snippets and social previews; 50–60 characters is the consensus sweet spot. When the length cap kicks in, the trimmer breaks on a word boundary rather than mid-word, so you do not end up with `introduction-to-typesc`.

Non-Latin scripts (Chinese, Arabic, Cyrillic, Devanagari) are stripped rather than transliterated. Romanization is a genuinely hard linguistics problem with multiple competing standards per language — pinyin vs. Wade-Giles for Chinese, ISO 233 vs. BGN/PCGN for Arabic — and getting it wrong silently produces wrong slugs. For multilingual sites the modern practice is to keep the source script and rely on percent-encoded UTF-8 URLs (`/blog/你好世界`), which browsers handle correctly and which read more naturally to native speakers than any transliteration.

The whole slugifier is a pure function under 2 KB. It runs synchronously in your browser as you type, makes no network requests, and has no dependencies. Verify in DevTools: typing in the input box produces zero network activity.