Is your site ready to be cited by AI?
Paste a URL for a one-shot check of AI crawler access, structured data, and citation readiness. No signup, no account — just a report.
What the audit checks
Seven signals, weighted by how much each one decides whether a model can read your page and attribute it to you. Every check links to a page explaining what it looks for and how to fix it.
- AI crawler access20 pts
Whether your robots.txt lets AI crawlers read your pages at all.
- Structured data20 pts
Whether pages carry valid schema.org JSON-LD describing what they are.
- Heading hierarchy10 pts
Whether headings form a real outline rather than styled text.
- Entity clarity15 pts
Whether a reader can tell who you are and what you do from the page alone.
- llms.txt10 pts
Whether you publish an llms.txt describing your site to AI agents.
- Readable content15 pts
How much of your page a crawler can read before JavaScript runs.
- Citation readiness10 pts
Whether the page carries marks of something worth citing: author, date, sources, and specifics.
Questions people ask first
- Is it actually free?
- Yes. No account, no card, no email: the report renders on the page as soon as it finishes. You can ask for a copy by email afterwards, and that is a separate opt-in you have to confirm.
- What do you keep about the site I audit?
- One aggregate row per run: the score, the seven pass/warn/fail verdicts, and a one-way hash of the domain. No URL, no IP address, no email, and nothing linking a row to whoever ran it. The hash exists only so that auditing one site twice doesn't count as two sites in the statistics we publish.
- Does it check my whole site?
- No. One page, plus robots.txt and llms.txt from the root of that domain. Point it at the page you actually want an assistant to cite. That is usually a product, pricing or documentation page.
- Does it run JavaScript?
- No, and that is deliberate: it reads the HTML your server returns, because that is what AI crawlers read. If your content only exists once a framework has booted, an empty page is the honest finding. It is a mistake we made on this site ourselves. Our own audit is what caught it.
- How will the audit appear in my server logs?
- As BrandtraceGeoAuditBot/1.0, with a link back to this page in the user agent. It requests the page you submitted, follows redirects, gives up after eight seconds, and reads nothing else besides robots.txt and llms.txt.
- Where do the seven checks come from?
- Published specifications. Crawler rules follow RFC 9309, structured data follows schema.org as Google documents it, and llms.txt follows the public proposal. The weighting between them is ours, and every check links to a page explaining what it looks for.
- Is 100 out of 100 the goal?
- 100 means a machine can read your page and work out who it belongs to. It does not mean models will talk about you. That depends on what you publish and on what everyone else says about you. Treat the score as a floor to clear.
- How often should I re-run it?
- Whenever robots.txt or your page templates change. Both get edited by people who are not thinking about AI crawlers that day, and both can undo a passing score in one commit. Ask for the re-check and we run it every 30 days, emailing you only when the score actually moves (in either direction).