AEO checker. Grade any page for AI search.

An AEO checker grades a page for answer engine optimization — how ready it is to be read, extracted and cited by AI assistants like ChatGPT, Claude, Perplexity and Google AI Overviews. This one fetches your page, your robots.txt and your llms.txt, then runs deterministic checks across five categories: AI crawler access, structured data, answer structure, snippet and meta signals, and trust signals. Every check shows what we found and the exact fix.

The score is 0–100, weighted across the five categories, with a letter grade: A is 85+, B is 70+, C is 55+, D is 40+, anything below is an F. Critical failures — AI crawlers blocked, noindex, or thin content — cap the score regardless of everything else, because they make the page invisible to answer engines no matter how good the rest is.

Most pages that have never been optimized for AI search land in the C–D range. The fastest wins are usually crawler access and structured data — they are binary fixes.

A grader is the instant snapshot: one page, one score, the prioritized fix list — what this tool does in about thirty seconds. An AEO audit is the deeper engagement: the same components reviewed across your whole site, plus the measurement side — which prompts you actually appear in. For the audit side of the equation, pair this grader with our free AI visibility checker, which runs live buyer prompts and shows where you are absent.

HubSpot’s AEO Grader measures brand perception: it asks AI models how they characterize your company and scores sentiment, brand recognition and share of voice. It never fetches your website — there is no crawl, no schema check, no robots.txt analysis.

This checker grades the opposite side: the actual page AI engines must read. Crawler access for 12 AI bots, llms.txt, JSON-LD types, answer structure, freshness. The two are complementary — HubSpot tells you how AI talks about you, this tells you whether AI can read and cite you at all.

At minimum: GPTBot and OAI-SearchBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot, and Google-Extended. We check 12 crawlers including CCBot (Common Crawl, the training set behind many models), Amazonbot, Applebot-Extended and Meta-ExternalAgent. Blocking a crawler removes you from that engine’s answers — sometimes intentional, but it should be a decision, not an accident left over from a 2023 block-all policy.

Structured data is the clearest machine-readable statement of what a page is, and pages with FAQ schema have been measured getting cited in AI answers at substantially higher rates than pages without it. Schema alone does not rescue weak content — but on a page with real answers, FAQPage, Article and Organization markup remove every excuse an engine has to misread you.

llms.txt is an emerging standard: a markdown index at your site root that hands LLMs a curated map of your most important pages. Adoption is early and no engine has committed to it as a ranking input, which is why we score it as a bonus rather than a penalty. It costs an hour to publish and it is one more surface where you control the narrative.

Free, no email, no signup, no result held hostage behind a form. Roman makes money when companies want the gaps fixed continuously — our agents write and publish answer-ready content built on what only your business can say. The grader is the honest demo.