Initial HTML contains substantive text

Many AI crawlers do not execute JavaScript. A single-page app that ships an empty shell and hydrates its content on the client is effectively blank to them, even when a browser-based renderer like Googlebot can see the full page. Server-rendering the main text is what makes a page readable to the widest set of agents.

page Content structure impl 1.0.0 html.ssr-content

How the check decides

The check reads the initial HTML response — the bytes the server returns before any client-side JavaScript runs — clones the <body>, removes <script>, <style>, <noscript>, and <template> content, and counts the words of visible text that remain. It passes at 50 words or more and fails below that threshold.

The 50-word floor is loose on purpose. Thin but legitimate pages — 404s, contact stubs, single-tile landings — sit well below 100 words and shouldn’t be penalised. What the threshold catches is the unambiguous failure mode: a body that’s only framework boot tags and an empty root div, where the actual text only materialises after hydration.

How to implement it

Render your main content on the server. Static site generators (Astro, Hugo, Jekyll, Eleventy) do this by default. For React/Vue/Svelte apps, use the framework’s SSR or static-export mode (Next.js, Nuxt, SvelteKit, Remix) so the first response contains real text rather than a <div id="root"></div> placeholder.

Pass

<body>
  <main>
    <h1>Install a14y</h1>
    <p>Run <code>npx a14y check &lt;your-site&gt;</code> to score any URL...</p>
  </main>
</body>

The text is in the initial response — no JavaScript required to read it.

Fail

<body>
  <div id="root"></div>
  <script src="/app.js"></script>
</body>

The content only exists after app.js runs, so a non-JS agent sees nothing.