The HTML <meta>
element represents any metadata information that cannot be represented by one of the other HTML meta-related elements (<base>
, <link>
, <script>
, <style>
or <title>
).
Depending on the attributes set, the kind of metadata can be one of the following:
- If
name
is set, it is document-level metadata, applying to the whole page. - If
http-equiv
is set, it is a pragma directive, i.e. information normally given by the web server about how the web page should be served. - If
charset
is set, it is a charset declaration, i.e. the charset used for the serialized form of the webpage. - If
itemprop
is set, it is user-defined metadata, transparent for the user-agent as the semantics of the metadata is user-specific.
Content categories | Metadata content. If the itemprop attribute is present: flow content, phrasing content. |
---|---|
Permitted content | None, it is an empty element. |
Tag omission | As it is a void element, the start tag must be present and the end tag must not be present |
Permitted parent elements | <meta charset> , <meta http-equiv> : a <head> element. If the http-equiv is not an encoding declaration, it can also be inside a <noscript> element, itself inside a <head> element. |
DOM interface | HTMLMetaElement |
Attributes
This element includes the global attributes.
Note that the global name
has a specific semantic in the <meta>
element and that the itemprop
must not be set when one of the name
, http-equiv
or charset
is already used.
charset
- This attribute declares the character encoding used of the page. It can be locally overridden using the lang attribute on any element. This attribute is a literal string and must be one of the preferred MIME names for a character encoding as defined by the IANA. Though the standard doesn't request a specific character encoding, it gives some recommendations:
- Authors are encouraged to use UTF-8.
- Authors should not use ASCII-incompatible encodings (i.e. those that don't map the 8-bit code points 0x20 to 0x7E to the Unicode 0x0020 to 0x007E code points) as these represent a security risk: browsers not supporting them may interpret benign content as HTML Elements. This is the case of at least the following charsets: JIS_C6226-1983, JIS_X0212-1990, HZ-GB-2312, JOHAB, the ISO-2022 family, and the EBCDIC family.
- Authors must not use CESU-8, UTF-7, BOCU-1 and SCSU, also falling in that category and not intended to be used on the web. Cross-scripting attacks with some of these encodings have been documented.
- Authors should not use UTF-32 because not all HTML5 encoding algorithms can distinguish it from UTF-16.
Notes:- The declared character set must match the one of the page. There is no valid reason to declare an inaccurate character set.
- This
<meta>
element must be inside the<head>
element and within the 1024 first bytes of the page, as some browsers only look at these first bytes before choosing a character set for the page. - This
<meta>
element is only a part of the algorithm to determine the character set of a page that browsers apply. The HTTP Content-Type header and any BOM elements have precedence over this element. - It is good practice, and strongly recommended, to define the character set using this attribute. If no character set is defined for a page, several cross-scripting techniques may be able to harm the page user, such as the UTF-7 fallback cross-scripting technique. Always setting this meta will protect against these risks.
- This
<meta>
element is a synonym for the pre-HTML5<meta http-equiv="Content-Type" content="text/html; charset=IANAcharset">
whereIANAcharset
corresponds of the value of the equivalentcharset
attribute. This syntax is still allowed, although obsolete and no longer recommended.
content
- This attribute gives the value associated with the
http-equiv
orname
attribute, depending on the context. http-equiv
- This enumerated attribute defines the pragma that can alter servers and user-agents behavior. The value of the pragma is defined using the
content
and can be one of the following:"content-language"
- This pragma defines the default language of the page.
Do not use this pragma, as it is obsolete. Use the global
lang
attribute on the<html>
element instead. "Content-Security-Policy"
- This value allows web site administrators to define content policies for served resources. With a few exceptions, policies mostly involve specifiying server origins and script endpoints. This helps guard against cross-site scripting attacks.
"content-type"
- This attribute defines the MIME type of the document, eventually followed by its character set. It follows the same syntax as the HTTP
content-type
entity-header field, but as it is inside an HTML Element, most values are not possible. Therefore the valid syntax for its content is the literal string 'text/html
' eventually followed by a character set with the following syntax:'; charset=
IANAcharset
' whereIANAcharset
is the preferred MIME name for a character set as defined by the IANA.Usage note:- Do not use this pragma as it is obsolete. Use the
charset
attribute on the<meta>
element instead. - As the
<meta>
may not be used to change the type of a document in an XHTML document, or in an HTML5 document following the XHTML syntax, never set MIME type to an XHTML MIME type that way. It would be incorrect. - Only an HTML document can use the content-type, so most of it is redundant: that's why it has been obsoleted and replaced by the
charset
attribute.
- Do not use this pragma as it is obsolete. Use the
"default-style"
- This pragma specifies the preferred stylesheet to be used on the page. The
content
attribute must contain the title of a<link>
element whosehref
attribute links to a CSS stylesheet, or the title of a<style>
element which contains a CSS stylesheet. "refresh"
- This pragma specifies:
- the number of seconds until the page should be reloaded, if the
content
attribute contains only a positive integer number; - the number of seconds until the page should be redirected to another, if the
content
attribute contains a positive integer number followed by the string ';url=
' and a valid URL.
- the number of seconds until the page should be reloaded, if the
"set-cookie"
- This pragma defines a cookie for the page. Its content must follow the syntax defined in the IETF HTTP Cookie Specification.
Note: Do not use this pragma as it is obsolete. Use the HTTP header set-cookie instead.
name
- This attribute defines the name of document-level metadata. It should not be set if one of the attributes
itemprop
,http-equiv
orcharset
is also set.
This document-level metadata name is associated with a value, contained by thecontent
attribute. The possible values for the name element are, with their associated value, stored via thecontent
attribute:application-name
, defining the name of the web application running in the webpage;Note:- Browsers may use this to identify the application. It is different from the
<title>
element, which usually consist of the application name but may also contain specific information like the document name or a status; - Simple webpages shouldn't define the application-name meta.
- Browsers may use this to identify the application. It is different from the
author
, defining, in a free format, the name of the author of the document;description
, containing a short and accurate summary of the content of the page. Several browsers, among them Firefox and Opera, use this meta as the default description of the page when bookmarked;generator
, containing, in a free format, the identifier to the software that generated the page;keywords
, containing, as strings separated by commas, relevant words associated with the content of the page;referrer
controlling the content of the HTTPReferer
HTTP header attached to any request sent from this document:Values for the content of <meta name="referrer"> no-referrer
Do not send a HTTP Referer
header.origin
Send the origin of the document. no-referrer-when-downgrade
Send the origin as referrer to a-priori as-much-secure destination (https->https), but doesn't send a referrer to a less secure destination (https->http). This is the default behavior. origin-when-crossorigin
Send a full URL (stripped from parameters) when performing a same-origin request, but only send the origin of the document for other cases. unsafe-URL
Send a full URL (stripped from parameters) when perfoming a a same-origin or cross-origin request. Note: Some browsers support keywords
always
,default
, andnever
for referrer. These values are deprecated.Note: Dynamically inserting<meta name="referrer">
(by document.write or appendChild) create a non-determinism when it comes to sending referrers or not. Note also that when several conflicting policies are defined, the No-referrer policy is applied.
The attribute may also have a value taken from the extended list defined on WHATWG Wiki MetaExtensions page. Although none have been formally accepted yet, a few commonly used names are among the proposals:
creator
, defining, in a free format, the name of the creator of the document. Note that it can be the name of the institution. If there are more than one, several<meta>
elements should be used;googlebot
, which is a synonym ofrobots
, but is only followed by Googlebot, the indexing crawler for Google;publisher
, defining, in a free format, the name of the publisher of the document. Note that it can be the name of the institution;robots
, defining the behavior that cooperative crawlers should have with the page. It is a comma-separated list of values from the following list:Values for the content of <meta name="robots"> Value Description Used by index
Allows the robot to index the page All noindex
Prevents the robot from indexing the page All follow
Allows the robot to follow the links on the page All nofollow
Prevents the robot from following the links on the page All noodp
Prevents the usage of the Open Directory Project description, if any, as the description of the page in the search engine results page noarchive
Prevents the search engine from caching the content of the page Google, Yahoo nosnippet
Prevents the display of any description of the page in the search engine results page Google noimageindex
Prevents this page from appearing as the referring page of an indexed image Google nocache
Synonym of noarchive
Bing Notes:- Only cooperative robots will follow the rules defined by the robots name. Do not expect to keep e-mail harvesters at bay with this.
- The robot still needs to access the page in order to read the meta value. If you want to keep them at bay, for example to prevent bandwidth consumption, use a robots.txt file instead (or in complement).
- If you want to remove the page of an index, changing the meta to
noindex
will work, but only when the robot visit the page again. Be sure not to prevent such visits, via the robots.txt file for example. Some search engines have developers tools, allowing a quick removal of some page. - Some possible values are mutually exclusive, like using
index
andnoindex
, orfollow
andnofollow
, at the same time. In these cases the behavior of the robot is undefined, and may vary from one to the other. So avoid these cases. - Some search engine crawler robots, like those of Google, Yahoo Search or Bing, support the same values on an HTTP directive,
X-Robot-Tags
: this allows them to use these pragma on non-HTML documents, like images.
slurp
, which is a synonym ofrobots
, but is only followed by Slurp, the indexing crawler for Yahoo Search;
Finally a few names are in common use, though not in the process of being standardized:
viewport
, which gives hints about the size of the initial size of the viewport. This pragma is used by several mobile devices only.Values for the content of <meta name="viewport">
Value Possible values Description width
a positive integer number or the literal device-width
defines the width, in pixels, of the viewport height
a positive integer number or the literal device-height
defines the height, in pixels, of the viewport initial-scale
a positive number between 0.0
and10.0
defines the ratio between the device width ( device-width
in portrait mode ordevice-height
in landscape mode) and the viewport size.maximum-scale
a positive number between 0.0
and10.0
defines the maximum value of the zoom; it must be greater or equal to the minimum-scale
or the behavior is indeterminate.minimum-scale
a positive number between 0.0
and10.0
defines the minimum value of the zoom; it must be smaller or equal to the maximum-scale
or the behavior is indeterminate.user-scalable
a boolean value ( yes
orno
)If set to no
, the user is not able to zoom in the webpage. Default value isyes
.Specification Status Comment CSS Device Adaptation
The definition of '<meta name="viewport">' in that specification.Working Draft Non-normatively describes the Viewport META element See also:@viewport
Notes:- Though not standardized, this attribute is used by different mobile browsers like Safari Mobile, Firefox for Mobile or Opera Mobile.
- The default values may change from one device, and browser, to another.
- To learn about this pragma in Firefox for Mobile, see this article.
scheme
- This attribute defines the scheme in which the metadata is described. A scheme is a context leading to the correct interpretations of the
content
value, like a format.Notes: Do not use this attribute as it is obsolete. There is no replacement for it as there was no real usage for it. Omit it altogether.
Examples
<!-- In HTML5 --> <meta charset="utf-8"> <!-- Defining the charset in HTML4 --> <!-- Note: This is invalid in HTML5 --> <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> <!-- Redirect page after 3 seconds --> <meta http-equiv="refresh" content="3;url=https://www.mozilla.org/">
Specifications
Specification | Status | Comment |
---|---|---|
Referrer Policy The definition of '<meta name="referrer">' in that specification. |
Working Draft | Defines values and semantics of <meta name="referrer"> . |
WHATWG HTML Living Standard The definition of '<meta>' in that specification. |
Living Standard | Added itemprop attribute |
HTML5 The definition of '<meta>' in that specification. |
Recommendation | Added charset attribute |
HTML 4.01 Specification The definition of '<meta>' in that specification. |
Recommendation |
Browser compatibility
Feature | Chrome | Firefox (Gecko) | Internet Explorer | Opera | Safari |
---|---|---|---|---|---|
Basic support | (Yes) | 1.0 (1.7 or earlier) | (Yes) | (Yes) | (Yes) |
name="referrer" |
17 | 36.0 (36.0) [1] | ? | ? | ? |
Constrained to values listed under referrer (as specified in the spec.) | 46.0 |
Feature | Android | Android Webview | Firefox Mobile (Gecko) | IE Mobile | Opera Mobile | Safari Mobile | Chrome for Android |
---|---|---|---|---|---|---|---|
Basic support | (Yes) | (Yes) | 1.0 (1.0) | (Yes) | (Yes) | (Yes) | (Yes) |
name="referrer" |
? | (Yes) | 36.0 (36.0) [1] | ? | ? | ? | (Yes) |
Constrained to values listed under referrer (as specified in the spec.) | No support | 46.0 | ? | ? | ? | ? | 46.0 |
[1] The referrer
value wasn't taken into account when navigation was happening via the context menu or middle click until Firefox 39.