In short. nofollow requests limits on following links, either across a page through robots rules or for an individual link through its rel attribute.

Definition: As of 17 October 2026, Google's robots meta tag specification defines nofollow as a page-level rule requesting that links on the page are not followed. The value also applies to individual HTML links through rel="nofollow".

Key facts

Google's robots specification supports nofollow in a robots meta tag and in an X-Robots-Tag HTTP response header.
Google's outbound link documentation recommends rel="nofollow" when other link qualifiers do not apply and the publisher wants to avoid association with, or crawling through to, the linked page.
Google's announcement about link attributes describes them as hints for ranking, with nofollow also becoming a crawling and indexing hint from 1 March 2020.

How does page-level nofollow work?

A robots meta tag is an HTML element that supplies crawler instructions for a page. Google's specification places it in the document's <head>. This example applies the rule to the page's links:

<meta name="robots" content="nofollow">

The name="robots" value addresses crawlers generally. Replacing it with name="googlebot" targets Google's search crawler. The specification treats the name and content attributes as case-insensitive.

X-Robots-Tag carries equivalent rules in an HTTP response header. Google's example targets its crawler explicitly:

X-Robots-Tag: googlebot: nofollow

The separate noindex rule excludes the page carrying it from search results. Google's combined example is:

<meta name="robots" content="noindex, nofollow">

The specification defines none as equivalent to noindex, nofollow.

How does link-level nofollow work?

An HTML link uses the <a> element, whose href attribute identifies its destination. The rel attribute describes the relationship to that destination. Google's outbound link documentation gives this example:

<a rel="nofollow" href="https://cheese.example.com/Appenzeller_cheese">Appenzeller</a>

Here the qualifier belongs to that individual link. For advertisements and paid placements, Google prefers rel="sponsored", while accepting nofollow. For user-generated content, such as comments and forum posts, it recommends rel="ugc".

These values can be combined. The same documentation shows rel="ugc nofollow" and also accepts comma-separated values. Regular links that Google should fetch and parse without qualifications require no rel attribute.

What should a website check?

An implementation review checks the intended scope: the page's robots meta tag or response header for page-wide instructions, and the individual anchor's rel value for a particular link. Google's robots specification explains that crawlers must access the page to read its meta tags or response-header rules. Blocking that page in robots.txt, the site's crawler access file, prevents Google from reading those rules.

Google's link guidance says qualified links are generally not followed, but their destinations can still be found through sitemaps or links from other sites and subsequently crawled. It recommends allowing crawling and applying noindex on a destination that needs exclusion from Google's index.

FAQ

Can I keep nofollow on existing paid links?

Yes. Google's link attribute announcement accepts existing nofollow links and recommends switching paid links to sponsored when convenient.

Can I combine nofollow with ugc?

Yes. Google's outbound link examples include rel="ugc nofollow" for a link in user-generated content.

Does page-level nofollow remove my page from search results?

The rule controls following links. Google's robots specification assigns exclusion from search results to noindex; its noindex, nofollow example combines both instructions.

Sources

Check how a page reads to answer engines. Open the GEO score analyzer