URL slug generator

Google's URL structure guide gives a few rules that apply directly to slugs:

• Readable words rather than long ID numbers.
• Hyphens, not underscores, between words. Google reads a hyphen as a word break.
• One case. Google treats URLs as case sensitive, so /Blog/My-Post and /blog/my-post are two URLs. Lower case avoids duplicates.
• Non-ASCII characters are percent-encoded when sent. Transliterating them keeps links readable everywhere.

Each line goes through the same steps, in this order:

1. Apostrophes inside words are removed, so don't becomes dont and not don-t.
2. &, @, % and + become and, at, percent and plus (switch off "Spell out" to drop them).
3. Letters are transliterated to ASCII: accents are stripped with Unicode NFKD normalization, and letters without a plain form are spelled out (table below).
4. Everything is lower-cased, and anything that is not a letter or digit becomes a word break. Emoji and symbols disappear.
5. Optional: numbers and stop words are removed.
6. Words are joined with the separator. With a max length, the slug stops at the last whole word that fits, and a cut slug never ends on a word like and or the.
7. In bulk mode with "Unique" on, a slug that already appeared gets -2, -3 and so on, still within the max length.

Empty lines stay empty, so line 12 of the output is always the slug of line 12 of the input. That makes it easy to paste a column from a spreadsheet and paste the slugs back next to it.

Most accented Latin letters are a base letter plus a combining mark, so Unicode normalization (normalize('NFKD')) splits é into e and an accent that can be dropped. Some letters have no such split and need a table:

Scripts without a Latin spelling here, such as Chinese, Japanese or Arabic, are dropped in ASCII mode. Turn off "Transliterate to ASCII" to keep them: the slug then works like Django's slugify(value, allow_unicode=True), and browsers send the letters percent-encoded. For the encoded form, see the URL encoder and decoder.

JavaScript, without a library:

function slugify(text) {
  return text
    .normalize('NFKD')                 // é -> e + accent
    .replace(/[\u0300-\u036f]/g, '')   // drop the accents
    .toLowerCase()
    .replace(/['’]/g, '')              // don't -> dont
    .replace(/[^a-z0-9]+/g, '-')       // everything else -> -
    .replace(/^-+|-+$/g, '');          // trim dashes
}

slugify('Crème Brûlée: 10 Tips'); // 'creme-brulee-10-tips'

// Or the slugify package (npm i slugify)
import slugify from 'slugify';
slugify('Crème Brûlée', { lower: true, strict: true }); // 'creme-brulee'

Python (Django):

from django.utils.text import slugify

slugify(" Joel is a slug ")                 # 'joel-is-a-slug'
slugify("你好 World", allow_unicode=True)    # '你好-world'

PHP:

// Laravel
use Illuminate\Support\Str;
Str::slug('Laravel 5 Framework', '-');   // 'laravel-5-framework'

// WordPress
sanitize_title('Crème Brûlée');          // 'creme-brulee'

Ruby on Rails:

"Donald E. Knuth".parameterize          # "donald-e-knuth"
"Donald E. Knuth".parameterize(separator: "_")  # "donald_e_knuth"

A new slug is a new URL. Links, bookmarks and search results still point to the old one, so add a permanent (301) redirect from the old URL to the new one. Google's redirect guide says a permanent server-side redirect is the best way to move a page and keep it in search.

Store the slug when the page is created instead of building it from the title on every request. Then fixing a typo in a title does not quietly change the URL. If two pages can share a title, make the slug unique in the database, the same way the "Unique" option adds -2.

Related free tools: the case converter for camelCase and snake_case names, the URL parser to check a full URL, and the meta tag generator for the rest of the page. For SEO products, see SEO tools on DevHunt.