Page metadata inspector
Check that a deployed page carries the metadata you intended before search engines and social platforms cache the wrong thing.
How it measures
- One GET from a single location reads up to 512 KiB of HTML; the document is scanned with bounded regular expressions, never rendered or executed.
- Findings follow widely published guidance: titles around 50 to 60 characters, descriptions up to about 160, a self-referencing canonical, one main heading, a viewport tag and a declared language.
- JSON-LD blocks are checked for JSON syntax and listed with their @type values.
Limits
- Client-side rendered content (metadata injected by JavaScript) is not visible to this check.
- JSON-LD is validated as JSON only, not against schema.org or any rich-result requirements.
- The inspection is heuristic: it points at likely problems, it does not predict rankings.
Troubleshooting
- Canonical points elsewhere
- Intentional for duplicate or parameterized URLs; a mistake when a template hard-codes the homepage canonical on every page.
- hreflang without self reference
- Each language version must list itself and every alternate, including x-default when you have a language selector.