GEO Audit Report
This page scores one homepage against 40 published checks across twelve weighted dimensions, and lists what failed with the evidence the check produced. The scan runs when the page loads.
Reading the homepage and robots.txt…
The scan fetches the page the way a crawler would, then reads robots.txt, llms.txt and the sitemap if they exist. It takes a few seconds because it does five real requests rather than reading a cache.
Nothing about the site is stored. The result exists only in the URL you are on, and re-loading this page runs the scan again.
| What is fetched | Why it affects the score |
|---|---|
| the homepage | Carries the highest weight of the twelve dimensions, because a page a crawler cannot read has no path to being quoted at all |
| /robots.txt | The site's stated intent about who may crawl, which is the only authoritative answer available to a scanner |
| /llms.txt | Worth 5% and no more, because the evidence behind it is weak |
| /sitemap.xml | Discoverability, and whether the site bothers to publish one |
What the score is measuring
Whether your pages are in a state that makes being cited possible. That is a narrower claim than it sounds, and a deliberate one.
Adding source citations produced the largest measured visibility gain for low-ranking sites, at +115%, ahead of the addition of expert quotations at +41% and statistics at +30-40%, across the strategies tested on generative engines. — Generative Engine Optimization, KDD 2024
Those figures are why citability and evidence carry 11% of the score and answer readiness a further 10%, while a context file carries 5%. Every weight, every check and every pass condition is published on the methodology page, and the analyser itself is open source at github.com/Alex13192/geo-scanner.
What this report cannot tell you
It does not ask any AI engine about you. Nothing here measures whether ChatGPT currently recommends your brand, and a high score is not a promise of a citation.
- Only the homepage is analysed. Interior pages are outside a single-URL scan entirely.
- A refused request is not proof of a block. Large sites verify crawlers by IP, so a scanner can be turned away while real crawler traffic is served normally. The verdict follows robots.txt, which is stated intent.
- Performance checks are coarse. Real Core Web Vitals need a browser, and this scanner deliberately does not run one.
Once you have a score, the badge generator turns it into something you can embed, and the llms.txt studio drafts the context file the AI Context Files dimension looks for.