Document prepared for this version of the website. Before public launch, it requires a legal assessment and confirmation of hosting configuration, log retention and deletion of working files. It does not constitute confirmation of the legality of the implementation.
1. One address, specific moment
The report is based on the public server response to the analyser’s request. It reads one HTML document, its headers and any redirects. The HTML module neither executes JavaScript nor logs in to the website. A server may return different content to a crawler than to a visitor’s browser.
Note the final address, date, criteria version, and missing information. The condition of the site, network, security and external resources may change the result. The measurement does not automatically describe all subpages of the domain.
2. Technical observations
We check selected elements such as title, meta description, headers, canonical, lang, viewport, links, image attributes, text available in HTML, main area and recognized JSON-LD types. The presence of an element does not prove the quality of its content or the compliance of the entire implementation with the standard.
For example, the presence of viewport does not prove that the phone is convenient to use, HTTPS does not prove the absence of vulnerabilities, and JSON-LD recognition does not guarantee an extended result in Google. The lack of a meta description is an observation, not proof that the contractor's contract is defective.
3. Robots, map and AI
A separate module downloads robots.txt and one XML map or map index from the same source. It does not go through addresses from the map. Read limits are up to 500 KiB for robots.txt and up to 512 KiB for map.
For Googlebot, bingbot, OAI-SearchBot, PerplexityBot and Claude-SearchBot we show matching rules for the examined address. GPTBot, ClaudeBot and Google-Extended are considered separately as content use declarations. Consciously blocking your training is not automatically an SEO error.
Interpreting robots.txt does not confirm an actual robot visit, passage through a firewall, indexing, or citation in AI. We do not read private Search Console settings. We do not give an aggregate "AI Ready" rating.
SEO/AEO/GEO Summary (Version 1.0) organizes existing readings: title, indexing and canonical; HTML text, H1 and main; JSON-LD syntax and types; rules of search robots. Shows source observations, missing data, and issues for manual verification. It is not an additional test or a point assessment of readiness for AI. Search and training remain separated. The availability of the text does not confirm its quality, and the presence of JSON-LD is not a condition for citation. The section is available in preview, JSON and PDF export.
4. Google PageSpeed: separate source
Upon explicit user selection, our server transmits the address to Google PageSpeed Insights. Google may perform a browser-based Lighthouse test. Phone and desktop results are separate, with 0-100 coming from the Google tool.
Lab metrics are not the same as every user's experience. The available CrUX data has a different scope and period; the report distinguishes data for address and origin. TBT is not a measurement of INP. Missing data, service limit and download error are not a zero score or a negative rating for the site.
The time it takes to download HTML from our server is not the time it takes to load the page in the browser. The result depends on the test conditions. Tool guidelines require review before implementation.
5. Configuration, security and availability
Reading HSTS headers, nosniff, CSP, embedding rules, and HTTP resource references is a passive configuration observation. We do not perform hacking, vulnerability scanning, account testing, form submission or purchase attempts.
The result is not an audit of security, digital accessibility, WCAG or regulatory compliance. The absence of a selected header does not in itself prove vulnerability; the presence of a header does not confirm that all security measures are effective.
6. How to interpret statuses
- Read correctly: a given criterion was met in the response obtained. This is not an assessment of the entire site.
- Signal to check: an automatic rule flagged a potential problem. Before changing it, you need to confirm its meaning and purpose of implementation.
- For verification: additional information or context is needed.
- No data / measurement unavailable: the module does not provide a basis for drawing a conclusion about the area under review.
The counts refer to control in the selected range. They do not constitute an assessment of the person, company, value for money or performance of the contract. A conscious noindex, canonical, or specific access rule may be a valid design element.
7. When the result is in doubt
Save the report, compare the address, date and scope with another study, then ask for a context assessment. Don't implement changes just to improve the color of the indicator. Fixing one element may impact site functionality or security.
The website owner and contractor can report an error or request clarification at [email protected]. The application does not require the ordering of BXM services. Detailed mode and permissions are described terms of use.
Have a question about the rules or the result? Write to [email protected]. Explaining an observation does not require ordering a paid service.