Can ChatGPT and Google's AI Read JavaScript-Rendered Content on My Webflow Site in 2026?

Here is the worry I hear from Webflow owners in 2026

A client called me last month a little panicked. She had read that ChatGPT and Perplexity cannot "see" modern websites, and she was sure her Webflow site was invisible to AI. I told her to take a breath. The truth is more calm and more useful than the scary headline.

I build on Webflow every day, and I care a lot about whether AI tools can actually read what my clients publish. So let me walk you through what really happens when an AI crawler visits your Webflow site, what it can read, and the one thing that can quietly cause trouble.

Do AI crawlers like ChatGPT and Perplexity run JavaScript?

No, they do not. The AI chat crawlers, GPTBot from OpenAI, ClaudeBot from Anthropic, and PerplexityBot from Perplexity, do not execute JavaScript. They read only the raw HTML that your server sends back in that first response. This is shown in 2025 analyses of AI crawler behavior, including Vercel's.

Think of it like this. When one of these bots asks for your page, it gets the plain text and tags that come back right away. It does not wait around, run scripts, or build the page the way your browser does. Whatever is not in that first batch of HTML simply does not exist to the crawler.

Does Google's AI read JavaScript differently?

Yes, and this is an important split. Googlebot does render JavaScript. It fetches your page, parses it, and executes the scripts, which is documented in Google Search Central. So Google Search and Google AI Overviews can see client-rendered content that the AI chat crawlers cannot. Gemini leans on that same Google pipeline.

This is why you cannot treat "AI" as one thing. Google's side of the world is patient and does the extra work of running your code. The chat crawlers are not. So a page that looks fine in a Google AI Overview might still be half-empty to ChatGPT or Claude.

Can AI crawlers read my normal Webflow content?

Yes, and this is the part that should calm you down. Published Webflow pages are served as static, pre-rendered HTML from Webflow's hosting. Webflow is not a client-side single-page app. So your normal text, headings, CMS-bound text on the page, and images are all present in that first HTML response.

That means the crawler reads your hero heading, your body copy, your section titles, and your CMS blog posts without any special trick. When people assume Webflow hides content behind JavaScript, they are picturing a React app, not Webflow. Webflow does the pre-rendering for you, which is a real gift for anyone who wants to be quoted by AI.

So what content is actually at risk of being invisible?

The risk is anything added by custom JavaScript or third-party embeds. Content injected purely by custom code runs in the browser after the page loads, so it is not in the initial HTML. If a bot never runs scripts, it never sees that text. This is the real gap, and it has nothing to do with Webflow's core.

Here are the usual suspects I run into. A pricing table pulled in by a third-party widget. A testimonial slider fed by an external script. An FAQ block loaded from another tool through an Embed element. A "reviews" section that a JavaScript app writes onto the page after load. All of that can read as blank to GPTBot and PerplexityBot.

Why is this a problem for getting cited by AI at all?

Because AI tools can only cite what they can read. If your best proof points, your clearest explanations, or your key answers live inside a JavaScript widget, the crawler grabs an empty shell. It cannot quote a sentence it never received. Your strongest content becomes your most hidden content.

This matters more now that people ask ChatGPT and Perplexity for recommendations instead of scrolling ten blue links. If you want the full story on how these systems decide which pages to pull from, I dug into that in my piece on how AI decides which pages to cite. The short version is simple. Readable text is the price of entry.

How do I check what an AI crawler sees on my page?

You have two easy self-audits, and neither needs a developer. First, open your live page and press Ctrl+U to view the page source. That is the raw HTML the crawler gets. Use Ctrl+F to search for a sentence you care about. If it is there, the bot can read it.

Second, open Chrome DevTools, disable JavaScript, and reload the page. If your key content still shows up, you are safe. If a section vanishes when JavaScript is off, that section depends on scripts, and the AI chat crawlers will miss it. I run this check on every site I care about, and it takes about two minutes.

What should I do if my important content fails the test?

Move that content into native Webflow. Rebuild the hidden section using real Webflow text elements, headings, and the CMS instead of a JavaScript widget. Once it lives in Webflow's own structure, it becomes part of the pre-rendered HTML, and every crawler can read it. That is the whole fix.

I am not saying rip out every embed. Some things, like a booking calendar or a chat window, are meant to be interactive and do not need to be quoted. The trick is to separate "content I want AI to cite" from "tools I want visitors to use." Keep the quotable stuff in native Webflow and let the interactive stuff stay as an embed.

Is custom code in Webflow always a bad idea then?

Not at all. Custom code is one of the reasons I love Webflow. You can add it through the Embed element, page-level custom code, and site-wide head or before-body-tag settings. The issue is never that custom code exists. The issue is relying on it to deliver text you want AI to read.

Here is my contrarian take after years of this work. Webflow is not the villain in the AI-readability story. Over-reliance on custom-code widgets and JavaScript embeds is. People blame the platform when the real habit to fix is stuffing important content into scripts. If you want to use custom code the smart way, I broke down safe patterns in my guide on adding custom code in Webflow. Use it for behavior, not for your core message.

How should I write my Webflow content so AI can actually use it?

Write it as plain, native Webflow text with clear headings and full sentences. Put your real answers in the body copy and your CMS, not in a popup or a script. Give each page enough substance to stand on its own. Crawlers reward clarity, and so do readers.

Length and structure matter too. Thin pages give AI almost nothing to grab, and padded pages bury the point. I wrote a full walkthrough on hitting the right depth in my post on writing content AI crawlers can use. Adding Schema.org markup as JSON-LD in your head code helps too, since that structured data sits right in the HTML where bots can read it.

Should Webflow owners actually panic about this in 2026?

No. If you build normal Webflow pages, most of your content is already readable by ChatGPT, Claude, Perplexity, and Google. The pre-rendered HTML does the heavy lifting. You only need to fix the specific spots where custom JavaScript hides text you want quoted, and now you know how to find them.

So do the two-minute audit, move any hidden content into native Webflow, and keep your embeds for the interactive things they are good at. That is the calm, honest version of the story, minus the scary headline. If you want a second set of eyes on your site, or you are not sure whether your key pages pass the test, reach out and I am happy to walk through it with you.