Special Characters in URLs: What to Avoid and Why
Spaces, ampersands, emoji, accents — special characters break or uglify URLs. Learn which characters are safe, which to avoid, and how to clean them up.

Type a space, an ampersand, or an emoji into a URL and something ugly happens: it either gets encoded into %-codes or breaks the link entirely. Understanding which characters are safe — and which to strip — keeps your URLs clean and reliable.
Why some characters are a problem
URLs were designed to use a limited, safe set of characters, defined in RFC 3986. Anything outside that set has to be percent-encoded into codes like %20 (a space) or %26 (an ampersand). The browser still works, but the URL becomes unreadable:
You wrote: /café & crème
Browser sees: /caf%C3%A9%20%26%20cr%C3%A8me
That string is impossible to read, hard to share, and looks broken in a search snippet.
Characters that are safe in a slug
For URL slugs, stick to this minimal, reliable set:
- Lowercase letters
a–z - Numbers
0–9 - Hyphens
-(as word separators)
That's it. This "alphanumeric + hyphen" set works everywhere, never needs encoding, and reads cleanly.
Characters to avoid (and what to do instead)
| Character | Problem | Fix |
|---|---|---|
| Space | Becomes %20 |
Replace with a hyphen |
& |
Reserved for query strings | Spell out as "and" |
= ? |
Reserved for query strings | Remove from the slug |
! ' " , . |
Punctuation clutter | Remove |
@ # $ % |
Reserved or encoded | Remove |
| Emoji 🎉 | Encoded into long % strings |
Remove |
| Accented letters (é, ñ) | Encoded unless transliterated | Convert to base letters |
| Uppercase | Case-sensitivity issues | Lowercase everything |
Spaces → hyphens
Never leave spaces in a slug. Replace them with hyphens, which search engines read as word separators. (See hyphens vs. underscores.)
Punctuation and symbols → removed
Apostrophes, exclamation marks, ampersands, and the rest add nothing to a URL and often break it. Strip them.
Mom's "Best" Pie & Cake Recipes! → /moms-best-pie-and-cake-recipes
Accents → base letters
Accented characters get percent-encoded into noise. Convert them to their plain form (é → e, ñ → n) so the slug stays readable. (See URL slugs for non-Latin languages.)
A special note on non-Latin scripts
There's an important exception. While accented Latin letters should be transliterated, fully non-Latin scripts like Arabic, Chinese, or Cyrillic can be used directly in modern URLs (these are called IRIs). An Arabic slug like /نصائح-السيو is valid and meaningful to Arabic readers — it shouldn't be transliterated into guesswork. The rule is: transliterate accents, but preserve real non-Latin scripts.
How to clean special characters automatically
Manually catching every symbol, accent, and stray space is tedious and easy to get wrong. A URL slug generator handles it in one pass: it lowercases, converts accents, removes punctuation and symbols, and swaps spaces for hyphens — while keeping legitimate non-Latin scripts intact.
The takeaway
Keep slugs to lowercase letters, numbers, and hyphens. Replace spaces with hyphens, strip punctuation and symbols, and convert accented letters to their base form — but preserve genuine non-Latin scripts. Clean characters mean readable, shareable, unbreakable URLs.
For the complete picture, see the SEO slug best practices.
Frequently asked questions
Which characters are reserved in URLs?
RFC 3986 reserves : / ? # [ ] @ ! $ & ' ( ) * + , ; = because they delimit parts of a URL. They must be percent-encoded when used as data, so keep them out of slugs.
Is a tilde (~) or period (.) allowed in a URL slug?
Both are "unreserved" characters and technically safe, but periods suggest file extensions and tildes look like legacy user directories. Stick to letters, numbers and hyphens.
Can I use emoji in URLs?
Technically yes — they are percent-encoded as long strings — but they are unreadable when pasted, hard to type and offer no SEO benefit. Remove them.