Free AEO Grader
Score any page on how easily an AI answer engine can quote it. 16 deterministic checks across answerability, structure, evidence, machine readability and crawler access — with the measurement behind every one, so you can see what to change rather than just a number.
Grade a single content page, not your homepage — the rubric scores whether one page answers one question well enough to be quoted.
What this grader is scoring
Ranking and being quoted are different competitions. A search engine picks ten pages and lets the user choose; an answer engine writes one answer and decides which passages to build it from. That second job rewards a specific shape of page: one that states its answer outright, breaks into self-contained chunks, attributes its claims, and is readable without running JavaScript.
Every check here is deterministic. Nothing is scored by a language model, so the same page gets the same grade every time and each point is traceable to something we measured.
The five categories
- Answerability (30 points) — headings phrased as real questions, an opening paragraph short enough to quote, and passages that stay under the length where retrieval starts truncating.
- Structure (20 points) — one H1, no skipped heading levels, at least one list or table, and enough depth to be worth citing over a rival page.
- Evidence and trust (20 points) — a named author, published and updated dates, outbound links to sources, and concrete figures rather than vague claims.
- Machine readability (20 points) — structured data that says what the page is, content present in the served HTML, a meta description, and a semantic wrapper with a canonical URL.
- Crawler access (10 points) — whether answer and indexing crawlers are allowed to fetch the page at all, and whether it is indexable once they do.
How to use this tool
- Grade one content page — a blog post or guide that answers a specific question, not your homepage. The rubric asks whether one page answers one question well.
- Fix crawler access first — if the answer crawlers are blocked, every other point on the page is theoretical.
- Then fix the opening paragraph — it is usually the cheapest change with the biggest effect, and most pages fail it.
- Re-grade after each change — the score is deterministic, so any movement is your edit rather than noise.
Common issues
- Burying the answer: an introduction that sets context for four paragraphs before answering means there is no passage to quote.
- Client-side rendering: the page looks complete in a browser and arrives empty to a crawler that does not run JavaScript.
- Undated content: without a published or updated date, an assistant treats the page as being of unknown age rather than as current.
- Headings written as labels:“Pricing” matches nothing a user typed; “How much does a widget cost?” matches the question directly.
- Claims without figures:“significantly faster” is not quotable, and a model has no reason to attribute it to you.
Frequently Asked Questions
What is AEO?
AEO — answer engine optimisation — is the work of making a page easy for an AI assistant to quote and attribute. It overlaps with SEO but is not the same job: classic SEO competes for a click, while AEO competes to be the passage a model lifts into its answer. GEO (generative engine optimisation) is another name for the same thing.
How is the score calculated?
16 deterministic checks across five categories: answerability (30 points), structure (20), evidence and trust (20), machine readability (20) and crawler access (10). Each check is pass, partial or fail, and the score is the points earned out of 100. No language model is involved, so the same page always gets the same score and you can see exactly which measurement moved it.
Why does the tool care about my first paragraph?
Because retrieval works on passages, not pages. An assistant looks for a self-contained span of text that answers the question outright, and a page that opens with three paragraphs of preamble gives it nothing to lift. Leading with a direct 40–60 word answer is the single highest-leverage change most pages can make.
Does a high score mean I will get cited?
No. This grades one page against the mechanics that make a page citable — it cannot know how strong the competing pages are, whether your brand is already known to the model, or which prompts your buyers actually type. Treat it as removing the reasons you would be skipped, not as a prediction. Measuring whether you are actually cited needs live prompts run against the assistants themselves.
Why did my JavaScript-heavy page score badly?
Most AI crawlers do not execute JavaScript. The tool measures how much visible text exists in the HTML your server sends, before any hydration. If a page renders its content client-side, a crawler sees an empty shell — so the content check fails even though the page looks complete in your browser.
Is this the same as an SEO audit?
No. There is deliberate overlap — structured data and crawlability matter to both — but this rubric ignores things that only affect ranking, like title tag length or internal link counts, and rewards things classic SEO does not score, like whether your headings are phrased as questions and whether claims carry quotable figures.
Enter your website to track its AI visibility across ChatGPT, Gemini, Claude, and Perplexity — and turn chatbot mentions into traffic.
- Set up in minutes
- Cancel anytime
Related Tools
See which AI crawlers your robots.txt lets through, and the rule behind each verdict.
AI Training Data CheckerSee whether your site appears in Common Crawl, the open corpus behind most AI models.
AI Visibility AuditRun live prompts against ChatGPT, Gemini and Perplexity to see if they mention you.
E-E-A-T CheckerAnalyse a page for experience, expertise, authoritativeness and trust signals.