How do I check whether an AI can read my site

The thirty-second version

1. 1Open the content settings.

   In Chrome: Settings, Privacy and security, Site settings, JavaScript.

2. 2Block JavaScript.

   Or add just your own domain as an exception, if you do not want to affect the rest of your browsing.

3. 3Reload the page.

   With Ctrl and R. Not from the browser cache.

4. 4Look at what is left.

   Headings, paragraphs, prices, opening hours, contact details. Or nothing.

The exact version, with a number at the end

The browser test gives you an impression. The command below gives you a number you can compare between sites and from one month to the next. It requests raw HTML, strips scripts and styles, removes the tags, and counts what is left.

curl -sL https://site-ul-tau.ro \
  | sed -e 's/<script[^>]*>.*\?<\/script>//g' -e 's/<[^>]*>/ /g' \
  | tr -s ' \n' ' ' | wc -c

It is not a laboratory measurement, but the order of magnitude is right and that is enough. Under a thousand characters, the page is practically empty to a machine. A few thousand, and it is readable.

The figures from three sites we built

We ran the test on three sites we built for clients, so as not to talk in the abstract. All three deliver the complete content without running a single line of JavaScript.

• cezic.ro

  2,730characters

  Services, descriptions, contact details.

  A static site, no application. The server sends the page already written.

• muzeulcfr.ro

  8,003characters

  Opening hours, address, history, collection, how to get there.

  Generated at publish time. All the text is in the HTML before it leaves for the browser.

• coralbaby.ro

  9,232characters

  Therapies, therapist qualifications, opening hours, prices, frequently asked questions.

  Generated at publish time, with the most useful content on the front page.

The gap between 2,730 and 9,232 is not a grade. cezic.ro is a brochure site with few services, while coralbaby.ro has therapies, qualifications, prices and frequently asked questions on the front page. Each delivers what it has. What matters is that it delivers, not how much.

If the page stays empty

It does not mean the site has to be rewritten. Almost every modern platform can deliver content from the server, the option is simply not on by default. In order of cost:

1. 1Turn on server-side rendering.

   Next.js, Nuxt, Astro, SvelteKit all do it. Usually it is a setting, not a rewrite.

2. 2Move the essential content into the HTML.

   Title, description, price, opening hours, contact. The rest can stay in the application.

3. 3Add structured data.

   They do not replace missing text, but they give the engine the exact facts instead of making it infer them.

4. 4Publish a static version of the pages that matter.

   The last resort, when the platform allows none of the first three.

What this test does not tell you

You can pass the test and still appear nowhere. The test measures one thing: whether the text is delivered. It does not tell you whether the bot is allowed in, which is a matter of robots.txt and the firewall. It does not tell you whether the text is good, whether it answers something anyone is actually asking, or whether someone else writes it more clearly than you.

It is the first step, not the only one. But it is first for mechanical reasons: as long as the text does not reach the engine, nothing you write afterwards matters.

Share this article

• Facebook
• X
• Threads
• WhatsApp
• LinkedIn