When a buyer asks ChatGPT or Gemini who to hire, the answer is assembled from what a machine could read on your website. If it could not read you, you are not ranked lower — you are simply not in the answer. These tools show you what it found.
Being found takes three things: a page a machine can read, a server that lets it through, and a page a phone can actually load. One instrument measures each. Point all three at your own domain before you point them at anyone else's.
Paste any URL. It fetches the page, parses it the way an answer engine does, and returns a grade with every deduction itemised. Nothing self-reported. About ten seconds.
Paste a URL. It measures what a phone has to download against what a laptop does, and how much of that is pixels no phone can display. Seconds, not adjectives.
Paste a URL. It grades what your server returns to anyone who asks — the thirteen things a corporate buyer checks before calling back, and the same layer that decides whether a crawler gets served at all. Plain language, no jargon.
The instruments tell you what is broken. These tell you why it matters, and what it costs while nobody fixes it.
They retrieve facts, they do not rank pages. What that changes about how you get found, what is measurable, and what to fix first.
URLs, hreflang and structured data written per language — and what breaks when a site is translated instead of built bilingual.
Per-item markup for objects that exist once, and the judgement underneath it: which provenance to publish and which to hold back.