Can ChatGPT see your site?

Paste a URL and see what AI crawlers get from it. The checker fetches your page without JavaScript, the way ChatGPT, Claude and Perplexity crawl, and compares it with the rendered page your visitors see.

What the checker does

For the URL you enter, the checker makes two requests and puts the results side by side:

The checker only reads the page. It does not store your page content, and it does not change anything on your site. It is free, with a limit on checks per hour so it stays available for everyone.

Why AI crawlers are tested without JavaScript: a study by Vercel and MERJ (December 2024) found that none of the major AI crawlers render JavaScript, while Googlebot does. Details and sources per crawler: Can ChatGPT read JavaScript?

How to read the result

The verdict compares the word count without JavaScript with the word count of the rendered page.

The cards show more than the verdict:

• Words: the visible text count. A large gap between the two cards means JavaScript adds most of your content.
• Headings: the headings found in each version. Headings that appear only in the rendered card are invisible to AI crawlers.
• Text excerpt: the first lines of text. A shell page often shows only the site name or "Enable JavaScript to run this app".

The crawler table lists common crawlers, whether they run JavaScript, and what each one would get from this page.

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

If the verdict is "Mostly empty"

Your page is built in the browser. People see it, Googlebot renders it, but AI crawlers read an almost empty HTML shell. This is typical for React, Vue or Vite single page apps and for sites made with Lovable, Bolt or v0.

You have three options:

1. Render on the server. Move the app to server-side rendering or static generation. This is the most thorough fix and usually means changing the framework setup. See the React guide and the single page application SEO guide.
2. Prerender for crawlers. Keep the app as it is and send crawler requests to a prerendering service that returns the rendered page. This is what Prebaked does.
3. Put the key text in the HTML. For a small site, adding the main text, title and description to the static HTML helps, but it has to be kept in sync with the app by hand.

With Prebaked, you add one snippet to your edge or server. It recognises crawler user agents and fetches the rendered page from Prebaked. People keep getting your site as before. Crawlers get the same page, with the content already in the HTML. Nothing is added or changed for bots. If Prebaked is slow or unavailable, the crawler gets your original page.

Install guides: Cloudflare Worker, Express, nginx, Netlify, Vercel. Already use another prerender service? Switch in a few minutes.

After installing, run the checker again. It tests your page the way a crawler without JavaScript fetches it, so a fix at the server or edge shows up in the result.

If the verdict is "Partly visible"

Part of the text comes from the server and part is added by JavaScript. Common causes are product lists, reviews, prices or tabs loaded after the page opens. Compare the two excerpts to see what is missing. If the missing part is what you want AI assistants to know, the same options apply as above.

What the checker does not tell you

• Whether an AI assistant will mention you. Being readable is a requirement, not a guarantee. Prebaked does not promise rankings or AI citations.
• Whether crawlers already visit you. Your server logs show that. Look for user agents such as GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot.
• Your whole site. The checker tests one URL. Check your most important page types (home, product, pricing, article) one by one.

More background: AI crawlers and JavaScript sites.

FAQ