Website Audit
RankX AI's technical crawl of your site. What it checks, how the score is weighted, and the rule that a check which did not run is never shown as passing.
The Website Audit crawls your site and reports technical problems: missing and duplicate metadata, broken links and resources, canonical mistakes, performance, accessibility of your content to a crawler, and sitemap coverage. It produces a score, a list of issues by severity, and the pages each one affects.
One rule shapes everything about it: a check that did not run is never rendered as a check that passed.
The three screens plus a reference
| Page | What it covers |
|---|---|
| Running an audit | Starting one, the modes, and reading the results |
| Page selection and cost | Which pages get crawled, and what that costs |
| What "a check did not run" means | The three audit paths and which checks each can assess |
| The issue reference | Every check, generated from the product |
Unlimited on every plan, priced per page
There is no monthly audit allowance to run out of on any plan. Audits are priced per page crawled, so a large site costs more than a small one, which is predictable and fair.
RankX AI still counts how many you run each month, because usage should be reportable, but the count never refuses one: unlimited here means never refused, rather than never measured.
A per-crawl cap stops a single audit consuming an unbounded number of credits, and the cap rises with the page-count band you choose. See page selection and cost.
Issues, warnings and weights
Every check belongs to a category and carries a severity and a weight:
- Categories: content, technical, and user experience.
- Severities: an issue is a problem; a warning is worth knowing and costs less; passed is a check that ran and found nothing wrong.
- Weights are how much a check costs the score when it fires.
The penalty is a ratio of the right denominator for that class of issue, not of the page count for everything. A per-page problem is a ratio of pages crawled. A sitemap problem is a ratio of the sitemap. A site-wide fact, like a missing robots.txt, takes its full weight when it fires rather than being diluted to nothing by the number of pages you crawled.
That last one is not a nicety. Diluting a site-wide finding by page count made a fabricated finding take full weight while every real per-page finding contributed almost nothing, and ten audits in a row scored identically as a result.
Not assessed is a real verdict
A check RankX AI could not run reports as not assessed. It is not a pass, and it is not a failure: it moves the score in neither direction and is counted separately so the gap is visible.
This exists because seven checks once rendered confident green assurances, "canonical tags are present", "no duplicate content detected", for checks that had never run. Nothing was wrong with the site and nothing was right about the report.
Three checks are honestly reported as assessable on no path at all, because the crawler returns no field for them anywhere. They stay in the taxonomy with "not assessed on any path" rather than being deleted, because a visible gap records something a silent absence loses.
JavaScript rendering
The audit can render JavaScript during the crawl on every plan except the entry Direct one. That matters for a site whose content only exists after script runs: without rendering, the crawl sees what a plain fetch sees.
It is also worth knowing that no AI crawler executes JavaScript, so a site that needs rendering to be readable is a site the AI assistants cannot read either. AI Readiness checks for exactly that, and it is the more consequential of the two findings.
What happens to the findings
Audit issues become candidates for the Tasks board, deduplicated so one page does not appear four times, and ordered by opportunity rather than by category.
On a connected WordPress site, some of them can be applied rather than copied out, with a dry run first. See the WordPress integration.
Where to go next
- Running an audit.
- What "a check did not run" means, which is the concept behind the whole report.
- The issue reference, every check with its weight and its coverage.
Why my traffic data is missing
The four states a Google connection can be in, only one of which returns figures, plus the two to three day publishing delay that is not a drop.
Running an audit
Starting a Website Audit in RankX AI, the modes available, how long it takes, and how to read the report when it finishes.