Skip to content
Fetching contributors…
Cannot retrieve contributors at this time
160 lines (126 sloc) 5.88 KB


Bleach is an HTML sanitizing library that escapes or strips markup and attributes based on a white list. Bleach can also linkify text safely, applying filters that Django's urlize filter cannot, and optionally setting rel attributes, even on links already in the text.

Bleach is intended for sanitizing text from untrusted sources. If you find yourself jumping through hoops to allow your site administrators to do lots of things, you're probably outside the use cases. Either trust those users, or don't.

Because it relies on html5lib, Bleach is as good as modern browsers at dealing with weird, quirky HTML fragments. And any of Bleach's methods will fix unbalanced or mis-nested tags.

The version on github is the most up-to-date and contains the latest bug fixes.

Basic Use

The simplest way to use Bleach is:

>>> import bleach

>>> bleach.clean('an <script>evil()</script> example')
u'an &lt;script&gt;evil()&lt;/script&gt; example'

>>> bleach.linkify('an url')
u'an <a href="" rel="nofollow"></a> url

>>> bleach.delinkify('a <a href="">link</a>')
u'a link'

NB: Bleach always returns a unicode object, whether you give it a bytestring or a unicode object, but Bleach does not attempt to detect incoming character encodings, and will assume UTF-8. If you are using a different character encoding, you should convert from a bytestring to unicode before passing the text to Bleach.

Customizing Bleach

clean(), linkify() and delinkify() can take several optional keyword arguments to customize their behavior.


bleach.clean() is the primary tool in Bleach. It uses html5lib to parse a document fragment into a tree and does the sanitization during tokenizing, which is incredibly powerful and has several advantages over regular expression-based sanitization.

A whitelist of HTML tags. Must be a list. Defaults to bleach.ALLOWED_TAGS.
A whitelist of HTML attributes. Either a list, in which case all attributes are allowed on all elements, or a dict, with tag names as keys and lists of allowed attributes as values ('*' is a wildcard key to allow an attribute on any tag). Or it is possible to pass a callable instead of a list that accepts name and value of attribute and returns True of False. Defaults to bleach.ALLOWED_ATTRIBUTES.
A whitelist of allowed CSS properties within a style attribute. (Note that style attributes are not allowed by default.) Must be a list. Defaults to [].
Strip disallowed HTML instead of escaping it. A boolean. Defaults to False.
Strip HTML comments. A boolean. Defaults to True.


bleach.linkify() turns things that look like URLs or (optionally) email addresses and turns them into links. It does this smartly, only looking in text nodes, and never within <a> tags.

There are options that affect output, and some of these are also applied to links already found in the text. These are designed to allow you to set attributes like rel="nofollow" or target, or push outgoing links through a redirection URL, and do this to links already in the text, as well.

Add rel="nofollow" to non-relative links (both created by linkify() and those already present in the text). Defaults to True.
A callable through which the href attribute of links (both created by linkify() and already present in the text) will be passed. Must accept a single argument and return a string.
A callable through which the text of links (only those created by linkify) will be passed. Must accept a single argument and return a string.
Do not create new links inside <pre> sections. Still follows nofollow. Defaults to False.
Linkify email addresses with mailto:. Defaults to False.
Set a target attribute on links. Like nofollow, if target is not None, will set the attribute on links already in the text, as well. Defaults to None.


bleach.delinkify() is basically the opposite of linkify(). It strips links out of text except, optionally, relative links, or links to domains you've whitelisted.

Allow links to the domains in this list. Set to None or an empty list to disallow all non-relative domains. See below for wildcards. Defaults to None.
Allow relative links (i.e. those with no hostname). Defaults to False.


To allow links to a domain and its subdomains, allow_domains accepts two types of wildcard arguments in domains:


Allow a single level of subdomain. This can be anywhere in the hostname, even the TLD. This allows you to, for example, allow links to example.*. * will match both and

>>> delinkify('<a href="">bar</a>', \
... allow_domains=['*.ex.*'])
u'<a href="">bar</a>'
>>> delinkify('<a href="">bar</a>', allow_domains=['*'])
u'<a href="">bar</a>

To allow any number of preceding subdomains, you can start a hostname with **. Note that unlike *, ** may only appear once, and only at the beginning of a hostname.

>>> delinkify('<a href="">t</a>', \
... allow_domains=['**'])
u'<a href="">t</a>'

If ** appears anywhere but the beginning of a hostname, delinkify will throw bleach.ValidationError (which is a ValueError subclass, for easy catching).

Jump to Line
Something went wrong with that request. Please try again.