SEO Metadata Checker guide
What the SEO Metadata Checker reads from a home page, how to judge title and description length, canonical links and Open Graph tags, and how to fix them.
What the checker reads
Search results and link previews are built from a handful of tags in the page <head>. The title becomes the clickable headline of a result and the label of the browser tab. The meta description is the usual source of the snippet under it, and the canonical link tells search engines which address to index when the same page answers under several URLs.
Open Graph and Twitter card tags do the same job for shared links: messengers and social platforms read them to draw the preview card. None of these tags change what the page shows a visitor, which is why broken metadata often goes unnoticed until somebody looks at a search result or pastes a link into a chat.
How to use the SEO Metadata Checker
- Open the SEO Metadata Checker and enter a domain such as
example.com. The input is a domain, not a URL: the request always goes to the home page over HTTPS. - Read the verdict and the findings. A missing title is critical; a title or description outside its range, a missing description and a missing or foreign canonical are warnings; missing Open Graph tags are a note.
- Check the table under the findings. It lists every tag that was found, including
twitter:card, the robots meta tag, the viewport and the charset, which are shown but not graded. - Copy the snippet of a finding into the page template, deploy, and run the check again from the permalink.
- Open the Full report tab for the same page with recommendations, or write complete tags with Record Builders.
Because nothing is crawled, one check covers one template. Run it once per template you care about by checking the page source of an article, a product or a category page by hand, or by building their tags with the builder and comparing.
The page must answer over HTTPS
https://<domain>/ cannot be fetched, there is nothing to read and the check stops with a single critical finding. A certificate problem is the usual cause; the TLS Checker shows it.Title and description length
The checker uses 30–60 characters for the title and 70–160 for the description. These ranges are display guidance, not rules of any standard. Search engines truncate by rendered pixel width, so a title of wide letters is cut earlier than one of narrow letters, and they often rewrite both from text on the page.
| Tag | Missing | Outside the range | Inside the range |
|---|---|---|---|
<title> | Critical | Warning, with the current title | Passed |
<meta name="description"> | Warning, with a snippet | Warning | Passed, with the text |
A short title wastes the most visible line of the result; a long one loses its end, which is often the brand. A good pattern names the page first and the site last. For the description, one or two plain sentences that say what the page offers work better than a list of keywords, which engines ignore and readers skip.
<title>Pricing and plans - Example</title>
<meta name="description" content="Compare the three Example plans, what each includes and how billing works." />
<link rel="canonical" href="https://example.com/pricing" />
<meta property="og:title" content="Pricing and plans - Example" />
<meta property="og:description" content="Compare the three Example plans and what each includes." />
<meta property="og:image" content="https://example.com/share/pricing.png" />
<meta name="twitter:card" content="summary_large_image" />Canonical links
The same page is often reachable under several addresses: with and without www, over HTTP and HTTPS, with tracking parameters or a trailing slash. A canonical link names the one to index, so ranking signals are not split. It is a hint that search engines usually follow, defined as the canonical link relation in RFC 6596.
The checker compares the canonical host with the domain you checked, ignoring a leading www. A canonical on another host is flagged because it hands the page's signals to a different site. That is right for a syndicated copy that credits the original and wrong everywhere else; the classic accident is a staging or CDN host left in a template after a migration.
Redirects and the canonical should agree. If http://example.com ends at https://www.example.com/ but the canonical says https://example.com/, engines get two different answers. The Redirect Checker guide explains how to make every variant end at one URL, and that URL is the one to put in the tag.
Open Graph and Twitter cards
Open Graph tags come from a convention that messaging apps and social platforms read for link previews. They are not part of any search standard, so their absence does not affect indexing, which is why the checker reports missing og:title, og:description or og:image as a note. A shared link without them falls back to the page title or to the bare URL.
og:imagemust be an absolute URL; a relative path is ignored by the platforms that read it.- Use an image the platforms can fetch without cookies or a login, served over HTTPS.
twitter:cardselects the card layout. Without Twitter-specific title and description tags, most readers fall back to the Open Graph values.- Preview caches are long-lived: after a fix, a platform may keep showing the old card until it fetches the page again.
Limits and related checks
The checker reads what the server delivers. Tags added by JavaScript after the page loads show as missing, and that result is meaningful: metadata that decides how a page appears in search belongs in the server-rendered head. The robots meta tag is shown so you can spot a stray noindex, but it is not graded, and the checker does not read robots.txt or a sitemap.
For the response itself, the HTTP Headers guide covers the headers that sit next to the HTML, including X-Robots-Tag. The Record Builders guide shows how to write meta tags and a robots.txt that match. The Domain Health guide puts metadata into the wider picture of DNS, TLS and redirects.
FAQ
My title and description are set by JavaScript and show as missing. What now?
XGM reads the delivered HTML without running scripts, which is the state a crawler indexes on its first pass. Whether and when a crawler renders your JavaScript later is not something you control. Put the tags that decide how a page appears in results into the server-rendered head; the same applies to a canonical link that a framework injects after hydration.
Are 30–60 and 70–160 characters hard limits?
No. They are display guidance: engines truncate by pixel width and often rewrite the snippet from the page text. A title or description outside the range is valid and indexed; the warning means the visible part is likely to be padded out or cut off.
Can I check a page other than the home page?
Not with this checker: the input is a domain and the request goes to the root page over HTTPS. Metadata usually comes from templates, so check the other templates in the page source, or build their tags with Record Builders and compare.
The canonical points to another host and is flagged. Is that always wrong?
No. It is correct for a syndicated copy that should credit the original, and wrong for an original page, where it hands the ranking signals to another URL. The comparison ignores a leading www, so example.com and www.example.com count as the same host.
Why are missing Open Graph tags only a note?
Open Graph controls link previews on social and chat platforms and does not affect search indexing. A link without it still works; it just shows the page title or the bare URL instead of a card with an image.
Does the checker look at robots.txt or the sitemap?
No. It shows the robots meta tag of the home page if there is one, without grading it, and reads nothing else. A noindex in that tag or an X-Robots-Tag header keeps the page out of results whatever the title says, so look at both when a page does not appear.