Paste a page's HTML and see what its head tags actually say — title, description, canonical, robots, viewport, Open Graph and Twitter. It parses with the browser's real HTML parser rather than regular expressions, which matters because head markup in the wild has attributes in any order, mixed quoting and comments containing tags, all of which break a pattern-matching extractor in ways that look like a missing tag. Each finding says what it costs rather than adding to a score: a missing title is fatal, a missing description just means Google writes its own, and an unintended noindex is the entire answer to why a page is not indexed.
| Tag | If missing |
|---|---|
| Title | Fatal — Google invents one |
| Description | Google writes the snippet |
| Canonical | Duplicate content risk |
| Viewport | Fails mobile usability |
| noindex present | The page will not be indexed |
| og:image | No preview when shared |
Because a browser cannot fetch another site's HTML without that site's permission, and almost none grant it. A tool that says it analyses a URL from your browser is really sending it to a server — which then has a log of every URL you have checked. Pasting from view-source keeps it entirely in the page.
Right-click the page and choose View Source, or add view-source: before the URL. Select all, copy, paste here. You only need the head, but the whole document is fine.
Because when it is there by accident it is the whole story — no other change will get the page indexed, and it is the single most common cause of a page mysteriously not appearing. A checklist that lists it alongside a missing og:site_name buries it.
Almost always because it is a relative URL. Platforms have no page context to resolve it against, so they ignore it and share the link with no preview, reporting nothing. It must be the full address including https.
No. Google has ignored it since 2009 and said so publicly. It is not harmful, just pointless.
Because a hundred-point audit implies everything matters equally, and it does not. These are the tags that change whether and how a page appears. Adding forty cosmetic items would make the report longer and less useful.
No. The parsing happens in the page.
The markup is pasted rather than fetched, and the reason is worth stating because most competing tools are quiet about it. A browser cannot request another origin's HTML without that origin's explicit permission, and essentially no website grants it — so a tool that claims to analyse a URL from the browser is either routing the request through a server, which then sees every URL you check, or it is not doing what it says. Pasting from view-source keeps the whole thing in the page. Parsing uses the browser's own HTML parser rather than regular expressions, which matters because real head markup contains attributes in any order, single and double quotes, self-closing and not, and comments containing tags — every one of which breaks a regex-based extractor in a way that looks like a missing tag. Both `name` and `property` are read, since Open Graph uses the latter and a parser looking only at `name` reports every og tag as missing. The checks themselves are deliberately short and each one says what it actually costs rather than contributing to a score. A hundred-point audit implies every item matters equally, when a missing title is fatal, a missing description merely means Google writes its own, and a missing `og:site_name` is cosmetic. `noindex` gets special treatment because when it is present unintentionally it is the entire answer — no other change to the page will get it indexed, and a checklist that lists it fourteenth alongside minor issues buries the one thing that matters. Open Graph images are checked for being absolute, since a relative URL is ignored silently by every platform: there is no page context for them to resolve it against, so the link simply shares without a preview and nothing reports why. Everything runs in the page.