Meta Tag Analyzer

How it actually works

URL mode does not contact the site directly, because a browser cannot read a cross-origin page, so the page is fetched as text through an ordered chain of public CORS gateways: AllOrigins in JSON form first, then AllOrigins raw, CodeTabs and cors.lol. A reply is accepted only when it is real content, and a gateway error or rate-limit page is rejected rather than parsed, so a dead gateway never masquerades as your site. Whichever gateway answered is named under the results.

Whatever HTML arrives, from a gateway or pasted straight into HTML mode, is handed to the browser DOMParser. The tool reads the document title, the charset declaration, and every meta element that carries a name or a property with a content value, storing each as a key and value pair; Open Graph tags and Twitter card tags land in the same list because they are just meta properties.

The essential-tag count is a fixed checklist: title, description, keywords, viewport, charset, robots, og:title, og:description, og:image and twitter:card. The stats bar reports how many of those are present and how many are missing; the values themselves are shown verbatim, and the whole set can be copied or downloaded as JSON.

What that means for your results

HTML mode makes the tool work with no network at all: paste View Source from any page, including one behind a login or one that blocks every public gateway, and the same parser reads its tags. That fallback keeps the tool useful when the URL path fails.

Because extraction is the browser DOM engine rather than a hand-written regex over markup, malformed and inconsistent tags are handled the way a real parser handles them, and the JSON export preserves exactly what was found.

Where it falls short — honestly

It reports what is in the markup and nothing more: there is no scoring, no per-tag grade and no written recommendation, and the earlier idea of Fix or Optimal verdicts was never implemented, so the page only counts essentials present versus missing.

The fetch is a single static request, so tags that JavaScript injects after load are invisible, and a page rendered entirely client-side may come back with an empty head. The public gateways are rate-limited and sometimes down; when none returns the page the tool says so plainly instead of guessing, and the fix is HTML mode.

The listing keeps the first value stored per key, so a duplicated tag name can shadow a later one.

A real end-to-end workflow

Paste a URL and analyse your own or a competitor page, or switch to HTML mode for anything the gateways cannot reach; read the missing-essential count first, since those are the tags most worth supplying.

To author the tags this tool reads, use Meta Tag Generator; to check whether a page is indexable rather than what its head contains, Google Index Checker looks at robots and noindex signals.