example.com

No internal links were sampled within the budget.

Without a canonical, duplicate addresses of the same page compete with each other.

No sampled page declared a canonical URL.

Only 19 words are present in the served HTML — the page most likely renders its content in the browser.

There is no description to measure.

Search engines will compose their own snippet from page text.

One top-level heading was found in the served HTML.

• Example Domain

Observed heading sequence on the homepage; visual styling is not judged here.

The homepage returned HTTP 200 to an ordinary GET request.

• HTTP 200

Single-language sites do not need them; this is not counted against the score.

The address that was actually served after redirects was checked.

• https://example.com/

Nothing to describe.

Neither the robots meta tag nor the X-Robots-Tag header blocks indexing.

0 same-origin links were found in the served HTML.

No JSON-LD was found on the homepage.

The html element declares the document language.

• en

Crawlers assume everything is allowed, which is usually fine but leaves no place to declare a sitemap.

• HTTP 404

robots.txt does not disallow ordinary web search crawlers here.

Neither robots.txt nor /sitemap.xml pointed at one.

0 URLs read from the sampled sitemap documents.

These control how a link renders when shared, not how it ranks.

14 characters. This is a recommendation about how results are rendered, not a ranking claim.

• Example Domain

The homepage declares a title element.

• Example Domain

Fewer than three pages were sampled, so uniqueness says nothing yet.

Mobile rendering depends on this declaration.

These pages are how both people and models establish who is behind a claim.

3 AI search crawlers are allowed by robots.txt. A CDN or WAF can still block them at request time — robots.txt is only the declared policy.

• OAI-SearchBot
• Claude-SearchBot
• PerplexityBot

Short, self-contained passages are the parts an assistant can quote without distorting them.

Assistants weigh statements they can attribute and date more highly than anonymous, undated text.

No gate wording was found in the served text.

Without Organization, Product, Article or similar JSON-LD, a machine has to guess who this page belongs to.

Nothing on the homepage states when the content was published or updated.

Language declarations should agree with each other.

An emerging convention, deliberately weighted low: extractable content matters far more than this file.

The response was the site itself, not a challenge or verification screen.

The served HTML carries 19 words. Assistants that fetch a URL without executing scripts will see an effectively empty page.

A citation needs one address that stays valid.

Whether model-training crawlers are allowed is a policy choice. It is reported here and has no effect on the score.

• GPTBot: allowed
• ClaudeBot: allowed
• Google-Extended: allowed
• CCBot: allowed

These fetch a page because a person asked for that page. Reported as a diagnostic and never scored.

• ChatGPT-User
• Claude-User
• Perplexity-User