GEO Checker

AI companies run several crawlers with different jobs, and robots.txt addresses each by its own name. The checker evaluates robots.txt for 26 crawlers and control tokens, kept in a dated registry compiled from each vendor’s own documentation.

Search and answer crawlers, and agents that open a page when a user asks, count toward the score: if they’re blocked, your content can’t be used as a source in answers. Blocking model-training crawlers is a separate choice that never lowers the GEO Readiness Score and doesn’t affect search access — Google-Extended, for example, has no effect on Google Search or AI Overviews.

This section also covers bot challenges (such as a Cloudflare check page), noindex and nosnippet directives, Bing’s noarchive, and content that only appears after JavaScript runs.

An answer engine first needs to know what a page is about. A title tag and H1 that name the same subject, subheadings that describe their sections, and text kept readable with lists, tables and short paragraphs all make that easier.

The checker reviews titles and H1s, the heading hierarchy, generic headings, thin and duplicate pages, how much of a page is navigation, whether the language is declared correctly, and pages made only of images. It doesn’t grade writing quality; it measures the structure that decides whether a machine can tell what the page says.

AI answers usually cite a single paragraph or list, not a whole page — so every section should make sense on its own.

Search and answer systems want to know who published the content they use. When your organization’s name, logo, official profiles (sameAs) and contact details are defined in structured data under one stable @id, every page can point to the same entity.

The checker merges the JSON-LD and microdata of all sampled pages into one knowledge graph, then checks whether the WebSite links to its publisher, articles to their author and publisher, and pages to the site — and flags dangling references, conflicting @ids and markup that only appears after JavaScript. Nothing that isn’t published is assumed.

A real, verifiable author and publisher behind a piece of content is important context for any system that may cite it. The checker looks at visible bylines, links to an on-site author profile, whether the author is defined as a Person, and how that person relates to the publishing organization.

It doesn’t score expertise and never suggests adding titles or credentials. What it checks is whether the information you already have is presented consistently to readers and machines: one identity per person, a real biography on the profile page, HTTPS, and a reachable privacy page.