Methodology
Methodology
Results describe observed signals within a limited scope. Review evidence, applicability and unknowns before making changes.
Summary
Run state describes resource collection (completed, partial or failed). Counts describe distinct evaluated checks, not finding cards. Unknown checks remain explicit; optional probes are informational.
- Issue found
- Needs review
- Passed
- Not applicable
- Not determined
- Information
Scope of this check
- Initial HTML only. JavaScript and browser rendering were not executed.
- Permission in robots.txt is not proof that a real crawler can pass every network or security control.
- A nearby link is not proof of factual accuracy or source authority.
- API and authorization checks are relevant only when the site offers those capabilities.
- llms.txt is optional and experimental. Its presence does not guarantee AI visibility.
- This comparison uses page observations. It does not measure AI rankings or explain citation causality.
- This is a point-in-time baseline, not a monitoring history or an experiment result.
- Wording checks are not available for this content language. Applicable structural checks are still shown.
Evidence
Direct observations report response data. Heuristics flag candidates for human review; they do not establish factual truth.
When wording analysis does not support the page’s content language, the report marks those checks as not determined.
Auto combines root URL context with page-level Article JSON-LD, Open Graph type and author/date signals. A layout article tag alone is insufficient. Conflicting signals stay unconfirmed; your selection takes priority.
| Choose a check for your task | Input and checks | Applicability, thresholds and limits |
|---|---|---|
| AI crawler access | robots.txt and page meta/header directives; evaluate matching crawler and path rules. | Policy depends on the intended bot access. Permission does not prove real bot retrieval. |
| Raw HTML readability | Initial HTML: main text, H1/hierarchy, canonical, lang and index directives. | A client shell candidate has fewer than 200 main-text characters and at least 3 external scripts. Browser comparison is outside scope. |
| Sitemap & discovery | robots sitemap declarations, fallback /sitemap.xml, canonical/index directives in sampled pages. | At most 3 declared sitemaps, 3 children per index and 8 page samples; lastmod older than 2 years is a review signal. No full-site orphan claim. |
| Structured data & entities | Initial JSON-LD: parse syntax, type-specific properties, URL/date/price/currency values and declared names. | Only recognized types have required-property rules. Name text matching is heuristic. No rich-result eligibility or factual correction is established. |
| Content clarity | Paragraph sentences, vague phrases, question headings and semantic lists/tables in initial HTML. | Wording supports en/vi. Long sentences exceed 30 whitespace-separated words. Dense-list candidates have at least 4 separators and 20 words. Concise first sentences have at most 30 words. Editorial review is required. |
| Claims & sources | Currency, percentage and magnitude claim patterns, vague attribution, nearby source markup, descriptive links and article author/date signals. | Wording supports en/vi. Nearby means within the same p/li/td block. Article provenance applies only to confirmed articles; unconfirmed type is unknown. No fact-check or source-authority assessment. |
| Brand identity | First-party meta, Open Graph, Organization/WebSite JSON-LD and visible navigation/contact links. | Different names or descriptions are review candidates; wording differences do not prove contradictory facts. No external model is queried. |
| Page comparison | 2–5 submitted pages measured separately: extracted main-text characters, headings, JSON-LD blocks and external links. | Collection is informational. Failed pages are unavailable, not zero. More text or markup is not an overall quality ranking. |
| Technical baseline | One initial HTML page plus conventional robots.txt and /sitemap.xml probes; six categories of observations. | Categories are observations, not overall page grades. Article date applicability follows page type. Use focused tools for detailed checks. |
| Agent interface discovery | Page HTML, robots rules, sitemap declarations and 8 conventional discovery URL probes. | Raw content below 200 characters needs review. OpenAPI, OAuth and experimental interfaces are optional; HTTP 200 alone does not validate interface semantics. |
| llms.txt check | Optional llms.txt/llms-full.txt: availability, UTF-8, Markdown H1, duplicate references, HTTP and same-origin robots rules. | At most 20 references checked, 256 KiB per file. Draft includes verified homepage; existing source descriptions require review. Optional experimental format, no citation guarantee. |