What the scan checks
The visibility scan reads a site the way an answer engine's crawler does: it fetches your pages without running JavaScript and reads the files crawlers read first. It takes a domain, needs no account, and shows the results on screen as it goes.
What it fetches
- Your homepage, then up to 15 pages in all: from your sitemap when it has one, otherwise from the links on your homepage
robots.txt, and up to 3 of the sitemaps it lists- Each fetch has 10 seconds and 3 MB, and the whole scan has 20 seconds. A page it did not reach in time is left out rather than reported as broken
It obeys your robots.txt for its own fetching. A page your rules disallow is listed as skipped, so
a page it did not look at never reads as a page with nothing wrong.
The five checks
Answer engines can reach you
Reads robots.txt for 13 AI crawlers:
GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-User, anthropic-ai,
PerplexityBot, Perplexity-User, Google-Extended, Applebot-Extended, CCBot, Bytespider,
meta-externalagent
| Result | When |
|---|---|
| Pass | No crawler is disallowed, and every sitemap robots.txt lists resolves. No robots.txt at all also passes: crawlers read that as allowed |
| Fail | Any of the 13 is disallowed, or robots.txt lists a sitemap that does not resolve |
| Not checked | robots.txt could not be read, for example a server error or a timeout |
Your pages exist without JavaScript
Counts the text each page carries in its HTML before any script runs, which is all most crawlers read.
| Result | When |
|---|---|
| Pass | Every page read has 600 characters of text or more |
| Warn | A page has fewer than 600 |
| Fail | A page has fewer than 200 |
Your structured data is present and valid
Looks for JSON-LD on every page read.
| Result | When |
|---|---|
| Pass | JSON-LD is present and parses. The types found are listed |
| Fail | JSON-LD on a page does not parse, or no page has any |
What your homepage says first
Quotes the first paragraph after your homepage's main heading. It is shown, never graded: whether that sentence answers the question a buyer would ask is yours to judge.
Your pages link to each other
Collects the links each page read makes to the other pages read, and weighs them: a link from a page that many others link to counts for more. Names any page that no other page read links to, and the pages most linked.
| Result | When |
|---|---|
| Pass | Every page read is linked from at least one other page read |
| Warn | A page read, other than the homepage, has no link from any other page read |
| Not checked | Fewer than 3 pages were read |
It never fails. The scan reads at most 15 pages, so a page with no links among them may still be linked from a page the scan did not read.
Results
Each check shows one of: pass, warn, fail, not checked (the scan could not look, and says why), or read (shown for you to judge). A result always names what it saw: the crawlers blocked, the pages and their character counts, the types found.
For what happens to the domain you enter, see Your domain and your data.