Frequently Asked Questions

What is AI visibility?

AI visibility is whether AI assistants — ChatGPT, Claude, Perplexity, Google AI Overviews — can crawl your website, read its content, and cite it correctly in an answer. It overlaps with SEO but is not the same thing: a page can rank well on Google and still be invisible to an AI assistant, because it is blocked by robots.txt, its content only exists after client-side JavaScript runs (which most AI crawlers never execute), or it lacks the structural signals — direct answers, clear entity data, dates, authorship — a model needs to lift a correct, attributable answer out of it.

Does llms.txt actually help?

Probably not much, at least not yet — and we say that plainly even though this tool checks for it. llms.txt is a proposed convention for pointing AI crawlers at a summary of your site, but there is little to no public evidence that publishing one changes whether a model cites you. We still check for it, and we still publish one ourselves, because it costs almost nothing to add and this tool would look foolish failing its own audit over it. Don’t treat it as a priority fix; spend your effort on crawler access and answer-readiness first — see our methodology for how each signal is actually weighted.

Which AI crawlers should I allow in robots.txt?

At minimum: GPTBot and OAI-SearchBot (OpenAI/ChatGPT), ClaudeBot and Claude-SearchBot (Anthropic/Claude), PerplexityBot (Perplexity), and Google-Extended (Gemini and AI Overviews grounding). CCBot feeds Common Crawl, an upstream corpus several other models train on, so it is worth allowing too unless you specifically want to opt out of AI training. Remember that robots.txt is only half the story: this tool’s live crawler-access check requests your homepage as each of these agents, because a CDN or WAF can silently block them even when robots.txt says they’re welcome.

What is GEO (generative engine optimization)?

GEO is the emerging practice of shaping a website so generative AI systems can find, understand and cite it — the AI-era counterpart to SEO. It covers the same ground this tool measures: making sure crawlers can reach your pages, that your content survives without JavaScript, that a machine can identify your entity and publisher, and that your writing is structured so a model can extract a correct, attributable answer rather than a garbled paraphrase. It is a newer, less settled discipline than SEO, and some of its proposed practices (like llms.txt) are still unproven — we try to be explicit about which parts are evidence-backed.

How is this different from a regular SEO audit?

A standard SEO tool checks whether Google’s crawler can index and rank you. This tool checks a narrower, newer question: can an AI assistant read your page at all, and is the writing shaped so a model can pull a correct answer out of it? The two overlap — clean HTML and fast pages help both — but a site can pass every classic SEO check and still be functionally invisible to ChatGPT if its content only renders client-side, or if a CDN’s bot-fight mode quietly blocks AI crawlers while letting Googlebot through untouched.

Is this tool really free, with no signup?

Yes. Enter a URL, get a full report — no account, no email address, no credit card. The report page is also permanent and public at a shareable URL, which is part of how this tool stays free: link back to your report instead of us needing to gate it. If you’d rather your site’s report weren’t listed, see /removal.

How often should I re-scan my site?

Re-scan any time you’ve changed something the report flagged — updated robots.txt, fixed a rendering issue, added structured data — to confirm the fix landed. Otherwise, once every few weeks is plenty; AI crawler behaviour and your site’s content don’t usually change fast enough to need more than that. Every report page has a re-scan button, subject to a light rate limit so the tool can’t be used to hammer any one site.

What do the grades (A–F) mean?

The score is 0–100 across five weighted pillars, mapped to a letter grade: A (90–100) is AI-ready, B (75–89) is mostly visible, C (60–74) is partially visible, D (40–59) is hard to cite, and F (0–39) is effectively invisible to AI assistants. The full breakdown of every signal, its weight, and how the grade bands were chosen is on the methodology page.

Does this check if AI actually mentions my brand?

Not directly, and we want to be upfront about that. This tool answers "can AI crawl and read my site" — a technical readiness check. "Does ChatGPT actually recommend my brand when someone asks" is a different question, usually answered by tools (like Ahrefs' AI Visibility Checker) that prompt live models and count citations, which costs real money per check. Every report here also shows three free, no-cost "off-site AI knowledge signals" — Wikidata, Wikipedia and Common Crawl presence — as an honest, if narrower, proxy for brand-level knowledge-graph presence, without ever calling a paid AI model API. For the real thing, we'd actually use both this tool and a brand-mention tool together: see the methodology page's off-site signals section for the full reasoning.

Why does JavaScript rendering matter for AI crawlers?

Most AI crawlers — GPTBot, ClaudeBot, PerplexityBot among them — fetch raw HTML and do not execute JavaScript the way a browser does. If your content is injected client-side (a common pattern in single-page apps), a crawler can receive a nearly empty page even though a human visitor sees a fully populated one. This tool’s Content Accessibility pillar runs a headless-browser comparison against the raw HTML response specifically to catch this gap, because it is one of the most common and most fixable reasons a site scores poorly.