How CrawlCheck fits together
In one sentence
How it fits together is the page that puts every CrawlCheck measurement in order as one chain: a crawler reaches the site, receives something, may quote it, an answer engine may cite it, a visitor arrives, and a lead lands with its source attached. Reach, read and quote are measured on every free scan; cite, visit and lead are kept as a record under a licence.
Every part of this product measures one link in the same chain: a crawler reaches your site, receives something, may quote it, an answer engine may cite it, somebody arrives, and a lead lands with a source attached. Each link below says what is measured, whether it is free, and which live page shows it working.
The chain, end to end
Read top to bottom. A break at any link makes every link after it unmeasurable, which is why the order matters more than the feature list.
Reach — Can a crawler get in at all?
measured on every free scan
Fifteen crawler identities ask for the same page from one address in about a second, robots.txt is resolved per agent rather than read as one file, and what the edge actually served each identity is recorded next to what the file promised. Two made-up paths per host say whether a server answers everything with a page.
the crawlers we ask as · robots.txt per AI crawler · a real report
Read — What did it actually receive?
measured on every free scan
The delivered bytes, not the rendered page: how much of the response is markup a reader will never see, the five heaviest inline blocks named individually, whether the words need JavaScript to exist, whether the edge served a stale copy, and whether the machine files resolve and parse.
the sections a report scores · llms.txt audit · entity map validator
Quote — Is there a fact it can safely repeat?
measured on every free scan
A lead that names the subject and defines it, paragraphs short enough to lift, question headings that are answered, FAQ markup that matches the visible text, and an entity graph whose every sameAs was fetched back to see whether it points home.
Cite — Did an answer engine say your name?
licensed
The citation watch asks the engines on a schedule and records the answer: a first citation, a top-three place gained or lost, a site that cited you last time and does not now, and wrong facts appearing about you. The source domain is read from the answer itself, never from the redirect wrapper around it.
Visit — Did anyone actually arrive?
licensed
A first-party beacon records arrivals by referring source and names AI assistants as what they are. A missing referrer is reported as unknown and never as direct, because browsers and app webviews strip referrers and absence is not evidence.
Lead and call — Did it turn into money?
licensed
One quote form on your site records first-touch and last-touch source beside the lead. When the source resolves to direct, the form asks the customer how they heard about you, and that declared answer is stored beside the measured one, never over the top of it.
What holds it together
Four things turn six measurements into one record rather than six reports.
Where it breaks, and what that costs
| Link | What breaks | What it costs |
|---|---|---|
| Reach | The edge refuses one identity and serves another | You are absent from answers that were never allowed to read you |
| Read | The page is mostly markup, or the words need JavaScript | A crawler that arrives gets no text worth quoting |
| Quote | No defining sentence, or an entity graph that points nowhere back | An engine can read you but cannot safely say who you are |
| Cite | Another site holds the citation you had | The loss is invisible until someone asks why the phone stopped |
| Visit | The arrival carries no referrer | The work gets filed as direct and nothing is learned |
| Lead and call | The lead is recorded with no source | You cannot tell which half of the spend produced it |
Free, and what the licence opens
Reach, read and quote are measured on every free scan, in full, with every finding and both exports and no account. The three links after them need a licence, because each one keeps a record over time rather than reading a page once.
7 days of full access to everything. $1.
The scan is free and stays free — every section, every finding, JSON and CSV, no account. One dollar opens everything the licence gates on one domain for seven days: the record kept over time, the dashboard, change receipts between any two dates, which crawlers really reached your server, the lead form, the side-by-side against a competitor, YouTube channel checks, and the API.
One charge, nothing renews. Day eight is a decision you make, not a charge you discover.
Open 7 days of full access — $1
The key is shown on the next page and is the account — there is no password. What the week builds stays readable on the report for that domain afterwards.