PAGELENS / METHODOLOGY
Our methodology
Observable checks, with visible limits.
PageLens analyzes source HTML, not a fully rendered browser page. The result describes the source received at the time of the request. A search crawler, signed-in visitor or JavaScript-rendered browser may receive different content.
How to read a report
| State | Meaning |
|---|---|
| Issue | A missing title, a noindex signal, or invalid JSON-LD JSON. Noindex can be intentional. |
| Review | A person should check whether the signal matters for this page. |
| Detected | A particular element or signal was found. Its accuracy and usefulness are not verified. |
| Context | An optional element, an unavailable measurement or an important limitation. |
SEO checks
We inspect title and description tags, H1/H2/H3 presence, an HTML canonical declaration, robots and googlebot meta directives, fetched X-Robots-Tag headers, viewport and language attributes, image alt attribute presence, link counts, Open Graph title and JSON-LD parsing.
Internal links are classified relative to the supplied or final URL host. HTML base elements and subdomain relationships are not resolved. We do not fetch linked destinations. Empty alt attributes are counted as present because decorative images may legitimately use them. This does not verify accessibility quality.
JSON-LD is checked for parseable JSON and declared types. It is not checked against schema requirements, Google rich-result rules or visible facts. Microdata and HTTP canonical Link headers are not evaluated.
GEO readiness signals
We look for readable source text, topic signposts, section headings, lists or tables, authorship and organization metadata, external links and explicit time elements. These are transparent structural signals, not validated predictors of AI inclusion.
The content extractor excludes scripts, styles, templates, navigation, headers, footers and explicit hidden or aria-hidden elements. It prefers main/article content. It does not evaluate CSS visibility. Approximate word tokens are reported as context; the 80-token review threshold is a tool heuristic, not a search-engine rule or a required content length.
Article and Person types can prompt an authorship check, but do not prove that an author is named or qualified. External links can be social accounts or irrelevant pages; we deliberately label this signal for manual review rather than treating all links as reliable citations.
Why there is no universal score
A contact page, product page and technical article have different purposes. Adding unnecessary words, schema or headings just to raise a score can make a page worse. PageLens displays counts of findings instead of pretending to calculate search rankings or a universal GEO score.
Not measured
Actual search indexing, robots.txt rules, keyword rankings, backlinks, semantic relevance, claim accuracy, rich-result eligibility, full accessibility, Core Web Vitals and AI recommendations are outside this audit. Fetch time measures the server request only; it is not a browser performance score.
Further verification
Use Google Search Console for your own indexing and search data, PageSpeed Insights for browser performance, and Rich Results Test for applicable structured data.
Google explains that foundational SEO practices apply to its AI features, without special AI schema or new AI text files being required. Read the official documentation. Other platforms can have different mechanisms.