html-sanitizer is a library aiming at handling, cleaning and sanitizing HTML sent by external users (who you cannot trust), allowing you to store it and display it safely. It has sensible defaults to provide a great developer experience while still being entirely configurable.
Internally, the sanitizer has a deep understanding of HTML: it parses the input and create a tree of DOMNode objects, which it uses to keep only the safe elements from the content. By using this technique, it is safe (it works with a strict whitelist), fast and easily extensible.
It also provides useful features such as the possibility to transform images or iframes URLs to HTTPS.
If you discover a security vulnerability within the sanitizer, please follow our disclosure procedure.
This library follows the same Backward Compatibility promise as the Symfony framework: https://symfony.com/doc/current/contributing/code/bc.html
Note: many classes in this library are either marked
@final
or@internal
.@internal
classes are excluded from any Backward Compatiblity promise (you should not use them in your code) whereas@final
classes can be used but should not be extended (use composition instead).
Many thanks to:
- The Open Web Application Security Project from which many of the tests of this library are extracted (more specifically from OWASP/java-html-sanitizer) ;
- Masterminds/html5-php which is a great HTML5 parser, used by default in this library ;
- tgalopin/html-sanitizer: from which this library is a hard-fork from.