Myrah

    AEO Grader

    Grade any page for answer engine optimization in about two seconds. Free, no signup, and every finding names the actual thing that is missing on the actual URL.

    What this AEO checker looks at

    Eight categories, weighted by how much each one actually determines whether a page can be cited. Crawler access carries the most weight because it is binary and total; sitemap coverage carries the least, because it affects discovery rather than whether the page can be used once found.

    AI crawler access
    Whether your robots.txt lets GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-User, PerplexityBot, Google-Extended and Bingbot read the page. This is the check that overrides every other one: a page an assistant is not allowed to fetch cannot be cited no matter how well it is written. It is also the failure people are least likely to know they have, because it is usually inherited from a template or a plugin default rather than chosen.
    Structured data
    Which schema.org types the page actually declares, read out of its JSON-LD rather than guessed at. We name the types we found and the specific one that is missing for the kind of page it is — and we count blocks that fail to parse separately, because invalid schema is ignored entirely rather than partially credited. A trailing comma in a JSON-LD block is the same as having no schema at all.
    llms.txt
    Whether the site publishes a machine-readable summary at /llms.txt, and whether it follows the format: an H1 business name, a blockquote summary, H2 sections, absolute links. This is the newest signal in the set and by far the least adopted, which is exactly why it is worth having.
    Sitemap coverage
    Whether a sitemap exists and whether this specific URL is in it. Plenty of sites have a sitemap that silently excludes the pages they most want cited, usually because of a generator filter that has outlived its purpose.
    Heading structure
    One h1 or several, and whether the outline skips levels. An extractor builds its understanding of a page from the heading hierarchy, so an h2 followed by an h4 loses the relationship between a section and its subsection.
    Answer-shaped content
    Whether headings are phrased as the questions they answer, and whether comparisons and sequences are in lists and tables rather than buried in paragraphs. People prompt assistants in questions; pages that mirror the question are markedly easier to quote as the answer.
    Entity clarity
    Whether it is unambiguous who publishes this. Title, meta description, and an Organization or LocalBusiness block with a sameAs array. sameAs is the part almost everyone omits, and it is the part that actually connects the brand on your site to the brand in a model's world knowledge.
    Citability
    Depth, a canonical URL, a date the page can be dated by, and whether it cites anything itself. Assistants prefer sources they can tell are current and that reference something beyond their own marketing.

    What your grade means

    The score is a weighted average across the eight categories, mapped to a letter: A at 90, B at 80, C at 70, D at 60, F below that. A category that did not apply to your page — FAQ schema on a page with no FAQ content — is dropped rather than counted against you.

    Expect a low grade the first time. Most sites land at D or F, including well-built ones with good traditional SEO, because these are new signals and almost nobody has addressed them deliberately yet. That is the useful part: the gap between a C and an A here is a short list of concrete, mostly one-afternoon changes, not a content strategy.

    Work the fix list top down. It is ordered by weight, so the first item is the one costing you the most. If crawler access is failing, fix nothing else until it passes — every other improvement on that page is unreachable until an assistant is allowed to read it.

    AEO tools and AI visibility tools are not the same thing

    The terms get used interchangeably and they should not be. They measure opposite ends of the same problem, and the order matters.

    An AEO tool — this one — crawls your site. It is deterministic, instant and free to run, and it answers a question about you: is this page legible to an answer engine? Schema, permissions, structure, entity signals. Nothing about it depends on what any model happens to say today.

    An AI visibility tool queries the models. It runs prompt sets across ChatGPT, Claude, Perplexity and Gemini and reports whether you were named, in what position, and against which competitors. It answers a question about the market: when buyers ask, who gets recommended? That is probabilistic, it costs real money per run, and it changes week to week.

    Start here. Fix what the grader finds, because measuring what assistants say about a page they are not permitted to read is measuring noise. Then move to tracking what the models actually say — and see our methodology for how that half is measured.

    AEO grader questions

    What is an AEO grader?
    An AEO grader audits a page against the things answer engines rely on when deciding what to cite: crawler permissions, structured data, entity signals, content shape and machine-readable summaries. Answer engine optimization is the discipline of being the source an AI answer is built from, rather than a blue link on a results page. This grader scores those signals and hands back a prioritized fix list.
    Is this an AEO checker or an AI visibility checker?
    An AEO checker. The distinction is real and worth keeping straight. This tool crawls your site and grades what it finds — it never queries a language model, so it tells you whether your page is READY to be cited. An AI visibility checker does the opposite: it asks ChatGPT, Claude, Perplexity and Gemini real questions and reports whether you were named. One measures your site; the other measures the answers. You want both, in that order, because there is no point measuring answers about a page an assistant is not allowed to read.
    Is it free, and do I need an account?
    It is free and there is no signup. Enter a URL and you get the grade. There is a limit of ten runs per day per visitor, which exists so the tool stays available rather than to push you toward an account.
    What makes this different from other AEO tools?
    Specificity. Most AEO tools in this category return a vague score and a button that books a demo. This one names the actual missing schema type on the actual URL, quotes the h1 it found, and tells you which named crawler your robots.txt is turning away. If a finding could not name something real that was read from your page, it would not be in the report.
    Do I need an AEO tool if I already do SEO?
    They are not substitutes. An SEO tool grades a page against how search engines rank results — keywords, links, position. An AEO tool grades it against how answer engines decide what to quote, which is a different set of signals: whether named AI crawlers are permitted, whether structured data identifies the entity, whether a single paragraph can be lifted out and still make sense. A page can rank on the first results page and still be unreadable to an assistant, usually because the answer only exists after JavaScript runs, or because robots.txt waves Googlebot through and turns OAI-SearchBot away. This grader looks only at the second set.
    Can I use it as an AEO tracker to watch changes over time?
    Not on its own. Each run is a point-in-time grade with a shareable link, so you can re-run after a change and compare the two by eye. Continuous tracking across engines over time is what the full Myrah product does — this grader is the free, instant, site-side half of that picture.
    Does it crawl my whole site?
    No. It grades the URL you give it, plus the three root files that govern the whole site: robots.txt, llms.txt and sitemap.xml. That is five requests, it takes a couple of seconds, and it respects your robots.txt while doing it. If robots.txt disallows the page, we stop and report that as the finding — because an assistant would be turned away the same way.
    My grade is bad. Is the grader wrong?
    Probably not. Most sites score D or F on a first run, including plenty of well-built ones, because these signals are new and almost nobody has done them deliberately yet. The bands are not tuned to flatter. Work down the fix list — crawler access first, then structured data and entity clarity — and re-run.
    What does answer engine optimization actually change?
    It changes whether a model can identify what your page is, who published it, and which question it answers — and whether it is permitted to read it at all. None of that replaces having something worth citing. It determines whether the thing you have written is legible to the systems that increasingly decide what buyers see.

    Know what the models say next.

    The grader tells you whether your page can be cited. Myrah tells you whether it is being cited — across ChatGPT, Claude, Gemini and Perplexity, tracked over time.

    See pricing