Documentation
Back
  • Connecting a site and reading your first audit
  • How Noverva reports numbers
  • Reading your audit
  • How opportunity scores work
  • The REST API
  • Connecting an AI client over MCP
  • Plans and limits

Connecting a site and reading your first audit

How Noverva crawls a site you connect, what it checks, and what it deliberately does not claim to know yet.

Connecting a site is the whole of setup. Enter a domain, and Noverva reads what your robots.txt allows, one page at a time, then tells you what it found.

What happens when you connect a domain

The domain is checked before anything is stored. Noverva resolves it and refuses addresses that point inside a private network, along with URLs carrying embedded credentials or non-website ports. This is not paranoia about your site — a crawler is a server-side fetcher, so an unchecked one can be pointed at infrastructure you cannot reach yourself.

Once the site is saved, a worker picks it up within a few seconds and begins:

  1. robots.txt first. A missing file is treated as permissive. An unreadable one is treated as restrictive, because a file we could not parse is not permission.
  2. Sitemaps, if any are declared, distinguishing a sitemap index from a sitemap.
  3. Pages, one at a time, with a delay between requests. Your robots.txt can ask us to slow down. It cannot ask us to speed up.

Every response is stored as fetched. Nothing is overwritten by a later run, so a finding can always be traced back to the page that produced it.

What the audit checks

Fourteen rules, each of which must explain itself. A finding always carries the rule that produced it, the evidence behind it, and what to do about it. There is no score out of a hundred, because a number you cannot argue with is a number you cannot trust.

Rules cover indexability, canonical URLs, titles and meta descriptions, heading structure, image alt text, internal links to broken URLs, orphan pages, and malformed structured data.

What a partial crawl means

If some pages could not be fetched, the run is recorded as partial and the pages that failed are excluded from every count. A crawl that reached 40 of 50 pages does not report on 50. This matters more than it sounds: the alternative is a coverage figure that quietly improves whenever the crawler has a bad day.

What Noverva does not know yet

Opportunity scores describe how your site covers a term — what it discusses, what it targets, and how reachable those pages are. They do not include search volume or difficulty, because no demand provider is connected. A high score means a clear gap in your own coverage, not a busy market.

Where a number is missing, Noverva says so rather than estimating it.