How AI search reads JavaScript sites
Updated
Google renders JavaScript, but most AI crawlers only read the HTML your server sends. Why client-rendered sites look empty to ChatGPT and Claude, and how to fix it.
Modern sites built with React, Vue or Svelte often send the browser a nearly empty HTML page plus a JavaScript bundle that builds the content. Browsers handle that fine. Crawlers are a different story.
Google renders, most AI crawlers don't
Googlebot runs a headless browser and renders JavaScript before indexing (Google describes this in its JavaScript SEO basics), though rendering can be delayed. AI crawlers mostly don't. An analysis published by Vercel in December 2024 of crawler traffic on its network found that the major AI crawlers, including OpenAI's and Anthropic's, fetched JavaScript files but did not execute them. None of these vendors documents rendering either way, so behaviour can change, but the safe assumption is that an AI crawler sees only the raw HTML response. (Apple's Applebot, like Googlebot, does render pages in a browser.)
What that looks like
A typical client-rendered homepage returns something like:
<body> <div id="root"></div> <script src="/assets/index-3f9a.js"></script> </body>
To a crawler that doesn't run JavaScript, that page has no product description, no prices and no headings. The only text it can use is the title and meta description.
How to check your own site
- Run
curl -s https://yoursite.com | lessand look for your headline and pricing in the output. - In Chrome, open DevTools, press Ctrl/Cmd+Shift+P, choose "Disable JavaScript" and reload.
- Use BotVisible's free scan, which counts the words in your raw HTML and flags JavaScript app shells.
How to fix it
- Server-side rendering (SSR): Next.js, Nuxt, SvelteKit, Remix and Astro can render pages on the server so the HTML arrives complete.
- Static generation / prerendering: for marketing pages that rarely change, build them to static HTML at deploy time.
- Minimum fallback: if you can't change the architecture yet, put a plain-text summary (what you do, who it's for, pricing, contact) in the initial HTML or a
<noscript>block, and make sure your title, meta description and JSON-LD are in the server response.
Fix the pages that matter for discovery first: homepage, pricing, product and docs landing pages. App screens behind a login don't need to be crawlable.