AX checker

Paste a page URL and see how readable it is for AI crawlers and agents. We fetch the raw HTML without JavaScript and score ten factors out of 100.

The ten factors

  • Fetchable without JavaScript (20 points)

    The page returns HTTP 200 with an HTML content type in the raw response, no JavaScript needed.

  • Title, meta description and one h1 (10 points)

    A title, a meta description and exactly one h1 tell an agent what the page is.

  • Answer block near the top (15 points)

    A paragraph of at least 40 words appears before any h3, list or table, so the answer is near the top.

  • JSON-LD structured data (10 points)

    A valid application/ld+json block gives machines structured facts.

  • Canonical link (5 points)

    A canonical link names the preferred URL.

  • Text-to-HTML ratio of 10% or more (10 points)

    At least 10% of the HTML is visible text, not markup and scripts.

  • Response under 1.5 s (10 points)

    The server responds in under 1.5 seconds.

  • Robots.txt allows verified crawlers (10 points)

    robots.txt does not block Googlebot, OAI-SearchBot or ClaudeBot for this path.

  • llms.txt exists (5 points)

    An llms.txt file exists at the site root.

  • No crawler errors in 7 days (5 points)

    Verified crawlers did not hit errors on this page in the last 7 days. The free tool assumes none; Glyphex checks your real agent traffic.

Frequently asked questions

What is an agent experience (AX) score?

It is a 0 to 100 score for how easily an AI crawler can fetch and understand a page. Most AI crawlers do not run JavaScript, so the score is based on the raw HTML response.

Does the checker store my page?

No. The tool fetches up to 200 KB of the page, scores it and returns the result. Nothing is saved. It is limited to 10 checks per minute per IP address.

Why does a page fail when it looks fine in my browser?

Browsers run JavaScript. If your content is rendered on the client, the raw HTML an AI crawler receives can be nearly empty. Server-rendering the main content fixes most failures.

Want to see which crawlers actually visit your site? Check the AI bots directory or the AI traffic checker.