Can ChatGPT read JavaScript?

No, not when it crawls your site. ChatGPT's crawlers download the HTML your server sends, but they do not run the JavaScript in it, so content that a script adds in the browser is not part of what they read. A study by Vercel and MERJ found that none of the major AI crawlers render JavaScript; Googlebot does.

This guide shows what each crawler is documented to do, where the vendors are silent, and how to check your own pages.

AI crawlers and JavaScript: bot by bot

"Runs JavaScript" below means: does the crawler execute the page's scripts before it reads the content? The table separates what the vendor documents from what published research measured.

What the vendors say:

• OpenAI documents the purpose of GPTBot, OAI-SearchBot and ChatGPT-User and how robots.txt applies to each. The page does not say whether they run JavaScript.
• Anthropic documents ClaudeBot, Claude-User and Claude-SearchBot and their robots.txt rules. JavaScript rendering is not mentioned.
• Perplexity documents PerplexityBot and Perplexity-User. JavaScript rendering is not mentioned.
• Google documents that Googlebot crawls a page, queues it for rendering, runs its JavaScript in headless Chromium and indexes the rendered HTML. The same page notes that not all bots can run JavaScript.

So for the AI crawlers, the "No" comes from measurement, not from a vendor statement. The research is from December 2024, and a vendor can change its crawler at any time. If a vendor documents JavaScript rendering later, this table will change.

What the research measured

Vercel and MERJ analysed AI crawler traffic across Vercel's network and published the results on 17 December 2024. The findings that matter here:

• None of the major AI crawlers they looked at rendered JavaScript. The study names OpenAI (OAI-SearchBot, ChatGPT-User, GPTBot), Anthropic (ClaudeBot), Meta (Meta-ExternalAgent), ByteDance (Bytespider) and Perplexity (PerplexityBot).
• ChatGPT's and Claude's crawlers do download JavaScript files. JavaScript files were 11.50 % of ChatGPT's fetches and 23.84 % of Claude's. They fetch the files but do not execute them.
• Two exceptions: Google's Gemini uses Googlebot's infrastructure, which renders JavaScript, and AppleBot renders JavaScript through a browser-based crawler.

Fetching is not rendering

A crawler that fetches your page gets the HTML response from your server. Rendering is the next step: a browser engine runs the scripts, which build the page in memory. Only a crawler that renders sees what the scripts produce.

For a client-rendered app (a React, Vue or Vite single page app, or a site built with Lovable, Bolt or v0), the HTML response is often a shell: a title, a script tag and an empty <div id="root">. The text, headings and links are added later by JavaScript. A crawler that does not render reads only the shell.

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

Pages that send their content in the HTML (static sites, server-side rendering, prerendering) look the same to both kinds of crawler.

What this means for your site

• Google can still index a client-rendered page, because Googlebot renders it. Google's own guide still recommends server-side rendering or prerendering, because not all bots can run JavaScript.
• ChatGPT, Claude and Perplexity read only your HTML. If your content exists only after JavaScript runs, these crawlers have little or nothing to use when they answer a question about your product.
• Blocking is a separate question. robots.txt decides whether a crawler may fetch a page. Rendering decides what it sees after fetching. A page can be allowed and still look empty.

Prebaked does not guarantee that an AI assistant mentions or cites your site. What it changes is that the crawler gets the same complete page your visitors see.

How to check your own pages

1. Use the free checker. It fetches your page without JavaScript, the way AI crawlers do, renders it in a browser, and compares the word counts and headings.
2. Compare by hand. Run curl https://www.<your-domain>/ and look at the response. If your main text is not in it, crawlers that do not render cannot see it.
3. Look at your server logs. User agents such as GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot show which AI crawlers already visit you.

See what ChatGPT sees on your site

How to make JavaScript content readable for AI crawlers

There are three common ways to put the content into the HTML response:

Prebaked is a prerendering service. A small snippet at your edge or server recognises crawler user agents and sends those requests to Prebaked, which returns the rendered page. People get your site as before, and crawlers get the same page with the content already in it. Nothing is added or changed for bots. If Prebaked is slow or unavailable, the crawler gets your original page, so it is never worse than today.

Install guides: Cloudflare Worker, Express, nginx, Netlify, Vercel. More on why AI crawlers matter: AI crawlers and JavaScript sites.

FAQ

See what ChatGPT sees on your site