Reference
What we check, and what each one means
Nineteen signals across three groups. Each one has a page explaining what it is, why an agent cares, and exactly how to fix it.
Readable
Can an agent get your content and understand it.
Content in the HTML
Whether your words survive a request that never runs JavaScript.
AI crawler access
Which assistants your robots.txt lets in — and the difference between two very different blocks.
Structured data
JSON-LD is how you state facts about yourself instead of hoping they're inferred.
Basic metadata
The title, description and canonical tag: the few lines that describe a page before anyone reads it.
llms.txt
A one-file map that tells an agent which of your pages are worth reading.
Sitemap
Discovery for the pages nothing links to.
Semantic structure
Landmarks are how an extractor tells your content from your chrome.
Key facts
Name, description, pricing, contact — the four things an agent needs to represent you.
Actionable
Can an agent do something on your site without a human.
Bot access
Whether a non-browser client gets your page or a challenge screen.
Contact reachability
An address an agent can act on, not a widget it can only look at.
Form submittability
Whether your enquiry form works without a browser running it.
Booking path
The conversion moment, and whether software can reach it.
API surface
The difference between an agent describing you and an agent using you.
MCP endpoint
How an agent operates your product instead of reading about it.
Agent manifest
Where you tell an agent what it may do, before it guesses.
Trusted
Is the response clean enough to rely on.
See where your own site stands
All nineteen checks against your URL, in about twenty seconds.