What does the GEO site scanner check?
It checks twenty-six things in four groups: whether AI and search crawlers are allowed to reach the page, whether the content can be read without running JavaScript, whether the business behind it can be identified, and how quickly it loads. Each result says why it matters and links to the fix, often one of our other free tools.
| Group | What is tested | Why it matters |
|---|---|---|
| Access | Page status and HTTPS; robots.txt rules for Googlebot, Bingbot, OAI-SearchBot, Claude-SearchBot, PerplexityBot, the user-requested agents and five training crawlers; noindex and nosnippet; sitemap | A page a crawler may not fetch, or a search engine may not index, cannot be quoted. Google says a page must be indexed and eligible for a snippet to appear as a link in AI Overviews or AI Mode. |
| Readability | Words of text in the HTML as delivered, title, description, headings, the length of the opening paragraph, language and canonical | Many AI crawlers read the HTML without running scripts, so content that only appears after JavaScript runs can be invisible to them. |
| Identity | Valid JSON-LD, an Organization or LocalBusiness record, sameAs profile links, contact details, About and Contact pages, share previews | Assistants describe a business from what they can confirm. Consistent, machine-readable identity makes it easier to match the site to its listings. |
| Speed and extras | Response time from our server, HTML size, llms.txt | Slow or very heavy pages use up crawl time. llms.txt is reported but not scored, because it is a proposed format and Google says no special files are needed. |
How is the score worked out?
Each check has a weight, from 1 for small things such as a missing language tag to 10 for blockers such as a page search engines may not crawl. A pass earns the full weight, a partial result 40 per cent of it, and a failure nothing. The score is the share of available weight earned, out of 100. Checks marked “for information”, such as whether you block training crawlers, are your choice and do not count. Blockers cap the score however well everything else does: a page that does not load at 15, a noindex directive at 30, a page that is empty without JavaScript or that Googlebot may not crawl at 35, a failing robots.txt at 40, and blocking every AI search crawler at 55.
The “Fix these first” list takes failures before partial results, heaviest first. On most sites, fixing those three changes the picture more than anything else.
Should I block AI training crawlers?
It is a business decision, and the scanner does not score it. Training crawlers and search crawlers are separate. OpenAI documents that disallowing GPTBot signals that content should not be used for training, while OAI-SearchBot controls whether a site can appear in ChatGPT search answers. Google-Extended controls use in Gemini and does not affect Google Search. You can block training and still be found in AI search; the robots.txt generator writes rules for exactly that.
What happens to the address I enter?
Our server fetches the public page you entered once, along with its robots.txt, llms.txt and sitemap, and sends the raw material back to your browser. The checks run in your browser, and the results are not saved. The fetch identifies itself as GEOExpertScanner/1.0. The scanner only reads public pages on standard web ports and refuses private and internal network addresses.