What the analyzer fetches and counts
The tool requests the page once, follows redirects and reads up to 2 MB of HTML within a 10 second limit. It then scans every a tag that has an href and resolves relative URLs against the final page address, honouring a base href if the page sets one. Links using mailto:, tel:, javascript: or a bare # fragment are not web pages, so they are dropped, and fragments are stripped from the URLs that remain.
Each surviving anchor becomes one row. The anchor text is the visible text inside the tag, cut at 120 characters. If the tag only wraps an image, that image's alt text stands in and is shown as [image: ...]. The page title appears under the total count so you can confirm the fetch landed on the page you meant.
How internal and external are decided
A link counts as internal when its hostname shares the organisational domain of the fetched page, so blog.example.com, www.example.com and example.com are internal to each other, and two-label public suffixes such as co.uk or com.au are recognised so that shop.example.co.uk stays internal to example.co.uk. Everything else is external and is grouped by hostname, minus any leading www., in the top linked domains panel, which lists up to 15 domains with counts.
The suffix list behind this check covers the common two-label endings rather than the full public suffix list, so an unusual country-code structure can occasionally classify a sibling subdomain as external. The rows themselves are still correct when that happens.
The findings list, item by item
The first line is always the summary: how many links were found and how they split between internal and external. Each line after that is either a warning you should act on or an informational note.
Image rows do not appear in the table because img tags are read only for the alt check. The Broken Link Checker is the tool that requests each image file and reports whether it loads.
- Links with no anchor text: a warning, because crawlers and screen readers cannot tell where they lead.
- Links opening a new tab without noopener: informational. Add rel="noopener" to be safe on older browsers.
- Images without an alt attribute: a warning counted over every img tag, including decorative ones, which should carry an empty alt="" rather than none at all.
- More than 300 links on the page: a warning that link credit is spread thin and visitors face a wall of choices.
- Links carrying nofollow, sponsored or ugc: informational, so you can check that paid and user-submitted links are marked and editorial ones are not.
Common problems the table reveals
Open the external tab and you will often find one domain linked dozens of times from a widget or footer. Those repeated links carry little weight individually and are the first place to look when an external count seems inflated. The internal tab shows the opposite problem: an important page with one link from a buried paragraph while the privacy policy gets fifty from every template.
Use the filter box with a path segment such as /blog/ to check that a section is reachable, or with a word such as "click" to find generic anchors like "click here" and "read more" that tell search engines nothing about the destination. Rewriting those to describe the target page is cheap and worth doing.
The table shows the first 500 filtered rows. The CSV download always holds the full list, one row per anchor, so a repeated link appears as many times as it occurs on the page.
Limits and related checks
The analyzer reports what the HTML says, not whether any link works. A link to a page that returns 404 looks identical to a working one here. The Broken Link Checker tests each destination. It does not crawl beyond the one page either.
Fetches go through the same address checks as every other DNSfinder tool, so private, loopback and cloud metadata addresses are refused, and a page that cannot be reached within the timeout returns an error rather than a partial result. Requests are rate limited per visitor.
The Open Graph Checker reads the title, description and social preview tags from the same page, and the Robots.txt Generator & Tester tells you whether a crawler may fetch the URL at all.