How to Check Meta Tags: Where You Look Changes the Answer

Three Places to Check, Each Showing Something Different

What to Check, in Order: 5 Steps

1. 1. View Source to see whether the tags are in the original HTML

   Open the source with Ctrl+U (Windows) or Command+Option+U (Mac), then search for <title, name="description", name="robots", rel="canonical", and og:. If canonical or robots aren't here, treat them as missing.

2. 2. List the rendered tags from the DevTools console

   Paste the code below into the console. It lists the title, meta, and canonical tags on the rendered page and whether each sits in <head> or <body>. Tags that differ from step 1 were added or changed by JavaScript.

3. 3. Compare with the HTML a social crawler gets

   Social crawlers read OGP. Fetch the page in a terminal with Facebook's crawler User-Agent (facebookexternalhit) to see the HTML that social networks receive. Some frameworks serve it differently from browsers.

4. 4. Use URL Inspection to see whether canonical reached Google

   In Search Console's URL Inspection, check the user-declared canonical. If you set one but this is empty, Google didn't receive it. If the Google-selected canonical is a different URL, Google chose another page as canonical.

5. 5. Run a checker for missing tags and lengths in one pass

   Finish with a checker. It reads the original HTML from the server, so it sees what step 1 sees; tags that exist only after rendering show as missing. Compare with step 2 to decide whether that's a real issue.

Step 2: code to paste in the console

[...document.querySelectorAll('title, meta[name], meta[property], link[rel="canonical"]')]
  .map((el) => ({
    where: document.head.contains(el) ? 'head' : 'body',
    tag: el.getAttribute('name') || el.getAttribute('property') || el.getAttribute('rel') || 'title',
    value: el.getAttribute('content') || el.getAttribute('href') || el.textContent,
  }))

Step 3: fetch as a social crawler

curl -s -A "facebookexternalhit/1.1" https://example.com/ | grep -oE '<(title|meta|link)[^>]*(og:|description|canonical|robots)[^>]*>|<title>[^<]*'

For reading curl output (and the trap of matching text inside scripts), see can AI crawlers read your SPA. For URL Inspection states, see Discovered – currently not indexed.

What Passes, and What Needs Fixing, Tag by Tag

Fix in order of what decides whether the page appears in search at all: robots (noindex) and canonical first, then title and description, then OGP and viewport. Even a clean title may not show as written — Google also uses headings, og:title, and other text to build the title link.

How to write each tag is covered in title tags, meta descriptions, canonical tags, and OGP tags. This article sticks to checking that they're present and actually delivered.

Pitfall 1: Meta Tags Landing in <body> (Measured on This Site)

Since Next.js 15.2, dynamically rendered pages start sending HTML without waiting for metadata, then stream the tags into the end of <body>. Only crawlers that can't run JavaScript get them in <head> as before. On October 23, 2026 we checked the five dynamically rendered pages on this site; only the shared audit result page had its tags in <body>. Here's what that page returned when fetched different ways.

We also opened the page in a browser and checked the finished page in DevTools: all four tags were still in <body>, not moved to <head>. The Next.js docs say they verified that bots which run JavaScript and inspect the full DOM, such as Googlebot, interpret this correctly. Google, on the other hand, states that canonical is only accepted in the <head>.

The call: there's no reason to leave canonical — which affects whether a page shows up at all — where two official sources disagree. The same day, we changed the Next.js config (htmlLimitedBots) so every request gets metadata in <head>. It only affects dynamic pages; our statically built articles and tool pages already had their tags in <head>.

If your site runs on Next.js, fetch the source with a browser User-Agent and with facebookexternalhit and compare. If <title> moves, streaming is happening.

Pitfall 2: Checkers Picking Up an SVG Icon's Title

SVG logos and icons often include a <title> element for screen readers. Browsers use the first <title> outside any SVG as the page title. Tools that search by tag name alone pick up the SVG's title too and mistake it for the page title.

Direbase's checker had this bug. It kept overwriting with the last <title> it found, so an SVG title in the body won. On October 23, 2026, we fed these three HTML patterns to the same Cloudflare Workers runtime as production, before and after the fix.

If a checker shows an unfamiliar short title — often a site or logo name — check the source for a <title> inside an SVG first. The original HTML from step 1 outranks any tool.

Pitfall 3: "It's in DevTools, So It's Fine"

On sites that add canonical or description with a tag manager or JavaScript, DevTools shows the right values — but the source HTML has none, or different ones. Social crawlers don't run JavaScript, so OGP added later never reaches the share preview.

Google says that if you can't set canonical in the source HTML, leave it out of the source and set it only with JavaScript. The worst case is the source and JavaScript saying different things. Make "steps 1 and 2 agree" your pass/fail line.

Related Tools

• Meta Tag Check — title, description, OGP, and canonical at once
• SEO Score Check — noindex warnings plus everything beyond meta tags
• Structured Data Check — validate your JSON-LD

Sources (checked October 23, 2026)

• Google Search Central — Meta tags Google supports (meta keywords unused)
• Google Search Central — Specify a canonical URL (accepted only in <head>)
• Google Search Central — Title links (sources Google uses)
• Search Console Help — URL Inspection tool (user-declared canonical)
• Next.js — generateMetadata (streaming metadata, since v15.2.0)