Enter a public URL and the analyzer fetches the page, then sorts every anchor into internal vs external and follow vs nofollow, with a roll-up of the most-linked domains and the most common anchor text.
Pro tip: Run it on key landing pages to confirm important internal links use descriptive, follow anchors.
Every anchor on the requested page is read from the HTML and classified along two axes. The first axis is destination: a link is internal when its hostname matches the page you submitted, and external when it points somewhere else. The second axis is the rel attribute: any anchor carrying rel="nofollow" is counted as nofollow, and everything else is treated as follow.
On top of those counts the tool builds two ranked lists. The top external domains list shows which outside sites you point to most often, which is useful for spotting an over-reliance on one source. The common anchor text list surfaces the phrases you reuse, so you can see whether your linking vocabulary is descriptive or stuck on "click here" and "read more".
Relative hrefs such as /pricing are resolved against the page address before anything is measured, so a link written as a path still counts as internal. Each resolved hostname is compared to the root host of the page; matches go to the internal bucket and the rest to external. The follow split is a literal check for the nofollow token inside the link's rel value, and target="_blank" is recorded separately so it does not affect the follow count.
The response also lists the first batch of individual links with their type and rel tag, so you can verify the totals rather than take them on trust.
Alongside the totals, the analyzer counts how many distinct external domains the page touches and ranks the twelve most-linked of them. A healthy editorial page usually spreads its outbound links across several relevant sources; a long tail of one domain repeated many times can read as an affiliate funnel or a paid placement that should be disclosed. The most-common-anchor list is the companion view: it tallies the exact anchor strings and shows the twelve you reuse most, which is the fastest way to notice that half your links say "here" or "this guide" instead of describing the destination.
The tool also surfaces a few plain-language recommendations based on what it found — for example, a prompt to review whether any followed external links should be marked nofollow or sponsored, a flag when a page has outbound links but no internal ones, and a note when the page references more than twenty-five different external domains.
Run the analyzer on the final, published URL rather than a staging or preview address, because the markup a server returns to an anonymous request is exactly what most crawlers see. If your internal count comes back at zero on a content page, that is almost always worth fixing first: a page with no internal links is harder for crawlers to place in your site's structure and passes no internal signals onward. When you spot a single anchor phrase dominating the list, vary it so the text describes where each link goes — that helps both readers and the search engines that read anchor text as a relevance hint.
Only the HTML returned for that one URL is parsed; the tool does not crawl deeper pages and does not run JavaScript, so links injected by a script after load will not appear. It reads markup, not link quality, so it cannot tell you whether a destination is authoritative or whether a link is broken — that is a separate job for a broken-link checker. The detailed row list is capped at the first fifty links for readability even though the totals count every anchor on the page. If a host blocks automated requests or serves different markup to bots, the counts will reflect what was returned at request time, not necessarily what a human browser sees.
You submit only a URL. The page is fetched from our edge as an anonymous visitor, parsed to produce the counts, and the result is returned to you — the markup is not stored or shared. Because the request is unauthenticated, the analyzer only works on pages that are publicly reachable.
Input: https://example.com/blog/seo-basics
Output: Total links 42 · Internal 28 · External 14 · Nofollow 3. Top external domain: developers.google.com (4 links). Most common anchor: "learn more" (5x).
Three nofollow links here are the outbound affiliate references — exactly as intended.
Input: https://example.com/promo
Output: Total links 6 · Internal 0 · External 6 · Nofollow 0, with a note that no internal links were found.
Zero internal links is a crawl-depth flag worth fixing with a few contextual links.
Input: https://example.com/best-tools
Output: Total links 35 · Internal 9 · External 26 · Nofollow 0. Top external domain: partner-shop.example (18 links), with a prompt to review whether followed external links should be sponsored.
Eighteen followed links to a single shop is the pattern the sponsored-link reminder is designed to catch.
No. The analyzer reads the HTML as it is delivered. Anchors that a framework or script inserts in the browser after the page loads are not present in that markup, so they are not counted.
Classification is based on hostname matching against the root host of the page you entered. A link to a different subdomain does not match the root host and is therefore reported as external.
Image links and icon links often have no text node, so their anchor text comes back blank. That is expected and usually points to navigation icons or logo links.
No. The fetch is unauthenticated, so it only works on publicly reachable pages. Gated or members-only URLs return whatever the server shows an anonymous visitor.
No. It classifies and counts links but does not request each destination, so it will not tell you if one returns a 404. Use the Website Broken Link Checker for that, since it probes each link's status.
They are unrelated. The follow split is based solely on the rel attribute, while target="_blank" controls whether a link opens in a new tab. The tool records target="_blank" separately so it never affects whether a link is counted as follow or nofollow.
The headline totals include every anchor on the page, but the detailed list is capped at the first fifty links so the output stays readable. The counts and ranked lists are still computed from the full set.
Our editorial and data-source notes explain how each check works and where estimates come from.