A meta tag analyzer and on-page SEO checker
This meta tag analyzer fetches any public web page and checks the tags search engines and social networks read from its <head>: the title and meta description (measured in pixels, the way Google trims them), robots directives in meta tags and the X-Robots-Tag header, the canonical link, viewport, charset, language, hreflang alternates, Open Graph and X (Twitter) card tags, favicons and structured data. As an SEO checker it also reads the page body: the H1 and heading order, the amount of text, image alt attributes and internal and external links. Every finding goes into one scored list, ordered so the problems that cost the most come first.
Use it to audit a page before launch, find out why a page dropped out of Google, check a client’s site or confirm what a CMS plugin writes. Pages that block automated visitors, need a login or live on your computer can be checked by pasting their HTML.
How to check the meta tags of a page
- Type or paste the page address into Page address, or switch to Paste HTML and paste the page source (optionally with its address so relative links resolve).
- Press Analyse tags. The page, redirects and share image are fetched; the checks run in your browser.
- Read the score and the prioritised list: High items stop a page from being indexed or shown properly, Low items are polish.
- Compare the Google result preview and Social share preview with what you intended, then open the tables for every tag, heading and link count.
- Fix the page, then analyse it again to confirm the score moves.
Meta Tag Analyzer features
Pixel-accurate title and snippet checks
The title is measured at 20 px Arial against about 580 px and the description at 14 px against about 920 px — the same method as the SERP Preview tool.
Robots in tags and headers
Reads robots, googlebot and other crawler meta tags plus the X-Robots-Tag response header, so a noindex set by the server is not missed.
Canonical checked against reality
Detects missing, duplicate, relative and misplaced canonicals, and checks the status of a canonical pointing elsewhere.
Share image really loaded
The og:image is downloaded and opened, so you see its true size, aspect ratio and file size — and whether it loads.
Structured data parsed
Every JSON-LD block is parsed; syntax errors are reported with line and column. Microdata and RDFa types are listed too.
On-page content
H1, heading outline and skipped levels, word count (scripts ignored), images without alt and link counts — the body checks of a website SEO checker.
What each check looks for
| Tag | Passes when | Why it matters |
|---|---|---|
| <title> | One title that fits in about 580 px | Usually the clickable headline in Google |
| meta description | One description, 70 characters to about 920 px | Often used as the snippet; affects clicks, not rankings |
| robots / X-Robots-Tag | No noindex, nofollow or nosnippet unless intended | A forgotten noindex removes the page from search |
| link rel="canonical" | One absolute URL in the head, pointing to this page or a working duplicate | Decides which address gets indexed |
| hreflang | Valid codes, absolute URLs, this page included | Shows the right language version to each searcher |
| og:* and twitter:card | Title, description and a 1200 × 630 px image | Controls how shared links look |
| JSON-LD | Every block parses and has @context and @type | A block with one stray comma is ignored entirely |
| h1–h6 | One H1 with text; levels in order | Tells readers and crawlers the topic and structure |
| img alt | Every image has an alt attribute | Image search and screen readers rely on it |
Worked example: a redirecting canonical
Suppose you enter shop.example.com/leather-wallets. The fetcher follows a 301 to https://www.shop.example.com/leather-wallets/, but the canonical still says https://shop.example.com/leather-wallets. The analyzer reports “The canonical points to an address that redirects” as a High problem — make the canonical the final address.
The same page’s title, “Handmade leather wallets and card holders”, is 41 characters and about 386 px wide at 20 px Arial, so it passes comfortably. An hreflang code written en_UK fails twice over — the separator must be a hyphen and the United Kingdom’s ISO code is GB — so it should be en-GB.
Limits worth knowing
- Scripts are not run, so tags and text a JavaScript framework adds after loading are not seen; Google renders scripts, so also use URL Inspection in Search Console.
- It cannot see rankings, backlinks or Core Web Vitals field data — those need Search Console or a crawler with its own index.
- The fetcher is an automated visitor on Cloudflare’s network. Sites with bot protection may refuse it (HTTP 403) — paste the HTML instead.
- hreflang must be reciprocal: every version lists every other. This tool sees one page, so it checks the set on that page only.
- Structured data is checked for syntax and types, not for Google’s rich-result rules. Build markup with the Schema Markup Generator and generate missing tags with the Meta Tag Generator.
Meta Tag Analyzer FAQ
How do I check the meta tags of a website?
Enter the page address and press Analyse tags. Every tag found is listed with a pass, warning or problem.
Is this a free SEO checker for a whole website?
It checks one page at a time, head and body. Check your key templates — home, category, article — to cover a site.
Why does my page say noindex when I can’t find it in the HTML?
It may come from the X-Robots-Tag HTTP header rather than a meta tag; the analyzer reads both and names the source. SEO plugins and staging settings often add it.
Is a missing meta description a problem?
Not for rankings — Google writes a snippet from the page text instead. A specific description is often used as the snippet and can earn more clicks, so it is a warning.
Why is my og:image reported as too small or not loading?
The size is real because the image is downloaded. Below 1200 × 630 px large previews look blurry; under 200 × 200 px Facebook rejects it. “Not loading” usually means a wrong path, a relative URL or a server blocking automated requests.
Can it check pages that need a login or are on localhost?
Not by address — the fetcher has no cookies and refuses private networks. Paste the page source into the Paste HTML tab instead.
Does the score match Google’s view of my page?
No tool can show that. The score summarises these checks: problems count fully, warnings half, notes not at all. Use it to find fixes, not as a ranking prediction.