What changed when we audited CiteMaybe

Two production scans covered the same homepage, methodology page, and privacy page. Warnings fell from nine to three after we corrected page metadata and checked the published source.

The same three URLs, scanned twice

On November 13, 2026 we audited the live CiteMaybe site, made changes, and submitted a fresh scan. Both scans reached the homepage, methodology page, and privacy page. The second scan used rule version 1.0.1. The changes included page-specific descriptions, explicit canonical URLs, and an update to how the audit retrieves an answer after a matching FAQ question.

Rule outcomes across the three production pages
OutcomeFirst scanFollow-up
Passes3945
Warnings93

The follow-up recorded zero failures and three untested checks. Each page had one untested browser-rendering check because this audit reads raw HTML. There were 48 tested findings out of 51 applicable findings. Nine other findings were inapplicable.

Inspect the audit evidence (JSON) for the page URLs, content hashes, and selected rule observations from the comparison.

What we could see in the source

In the first scan, all three pages had a description, but the description was identical. Each page also lacked a canonical link. The follow-up showed a distinct description and a self-referencing canonical on each page. Those two checks account for the six warnings that cleared.

For example, the methodology page’s declared canonical resolved to https://citemaybe.com/methodology. Its live description at the time gave readers a concise summary of the method.

See how CiteMaybe chooses pages, reads crawl rules, matches questions and calculates rule coverage from raw HTML.

We also checked the answer returned for “Which pages does CiteMaybe check?” The FAQ on the homepage contained the sentence below, and the follow-up located the passage after the question.

The scanner starts at the URL you submit.

That is a checkable retrieval result. A reader still has to decide whether the passage answers everything they needed to know. How question matching works

One editorial question stayed open

All three remaining warnings came from content.identity, the organization or offering description rule. For each page the audit looked for a named offering in supported JSON-LD or a recognizable English introduction. The three pages missed that pattern in this snapshot.

The rule can miss a clear introduction written another way. For the homepage, a useful next edit is to name the product, explain what it checks, and identify who uses it. For a privacy notice, first consider whether another product introduction would help a reader understand the data handling. A warning is a reason to inspect the page in context.

Read this result in context

These figures describe a historical three-page sample of citemaybe.com. They measure the scanner’s recorded rule outcomes. No external answer or citation was measured in this comparison. The site and its content have since changed, so a new scan can select different pages and return different findings.

To inspect today’s source, open the homepage, methodology, or privacy page. To make your own comparison, save the first report, record the final page URLs and rule version, and repeat after the new HTML is live.

Start a fresh audit