How AI Coding Tools Are Destroying Your Visibility in ChatGPT

Q: Why doesn't ChatGPT cite my site?

ChatGPT retrieves information from its training data and, in browsing mode, from real-time web content. If your site uses client-side rendering (React SPAs, Vue apps), the raw HTML contains no content—just an empty div and a JavaScript bundle. LLM crawlers don't execute JavaScript, so they see nothing to cite.

Q: Do SPAs hurt LLM indexing?

Pure SPAs (single-page applications) with client-side rendering are invisible to most LLM crawlers. The HTML document contains no substantive content—just JavaScript references. To be cited by AI engines, you need server-rendered or statically-generated content.

Q: Can I use Lovable/Cursor and still rank in ChatGPT?

Yes, but not with default settings. Tools like Lovable and Cursor generate React SPAs by default. You need to add a server-rendered layer: deploy FAQ pages statically, add pre-rendering for key routes, or use a hybrid architecture that serves critical content as static HTML.

Q: Is Next.js bad for AEO?

No—Next.js is actually one of the better choices because it supports Server-Side Rendering (SSR) and Static Site Generation (SSG) out of the box. The problem arises when developers use Next.js in pure client-side mode or don't configure static generation for key pages.

Q: How do I fix AI visibility quickly?

Three immediate actions: 1) Identify your highest-value pages and implement static generation or SSR for them. 2) Add structured data (JSON-LD schemas) for FAQs, products, and articles. 3) Ensure your sitemap.xml lists all important pages and is accessible to crawlers.

Q: What is LLM retrievability?

Retrievability is whether an LLM can find, fetch, and extract information from your content during inference. It depends on: whether your pages are in the LLM's training data, whether real-time retrieval (RAG) can access your content, and whether that content is parseable from raw HTML.

Executive Summary: Most sites built with AI coding tools (Cursor, Lovable, v0, Replit, Framer) are completely invisible to ChatGPT, Gemini, Claude, and Perplexity. These tools generate React SPAs that render content client-side—the raw HTML is empty. LLM crawlers don't execute JavaScript, so they see nothing to cite. This guide explains the technical problem and provides the 4R Framework to fix it.

Key Facts

The Problem: AI coding tools generate React SPAs with client-side rendering
Why It Matters: LLM crawlers don't execute JavaScript—they see empty HTML
Who's Affected: Startups and teams building with modern AI-assisted development
Revenue Impact: Zero visibility = zero demand capture from AI discovery
The Fix: The 4R Framework—Rendering, Relevance, Retrieval, Recommendations
Time to Results: Days for real-time systems, months for training-data LLMs

AI Coding Tools and LLM Risk

Tool	Default Output	LLM Risk
Cursor	React SPA	High
Lovable	React + Vite SPA	High
Framer	Hybrid (better)	Medium
v0 by Vercel	React components	High
Replit	Varies	Medium-High
Gamma	Presentation-first	High
Webflow AI	Hybrid	Medium

The Technical Problem

When you run curl on a typical SPA, here's what you get:

<!DOCTYPE html>
<html>
<head><title>Loading...</title></head>
<body>
  <div id="root"></div>
  <script src="/bundle.js"></script>
</body>
</html>

# LLM extractable content: ZERO

LLM retrievers fetch raw HTML without executing JavaScript. If your content only exists after JS execution, it's invisible to AI answer engines.

The 4R Framework (Bull AI)

1. Rendering

Pre-render critical content server-side. Implement SSR or SSG for key pages. Ensure raw HTML contains full text, not empty divs.

2. Relevance

Add semantic structure and entity grounding. Use JSON-LD schemas (FAQPage, Product, Article). Implement proper heading hierarchy.

3. Retrieval

Optimize crawl surface. Ensure sitemap.xml is complete and accessible. Implement internal linking. Don't gate content behind login.

4. Recommendations

Test and monitor visibility. Run prompts against LLMs to verify citation. Track weekly. Adjust based on data.

How to Test Your Site Right Now

Run this in your terminal:

# What does the raw HTML contain?
curl -s https://your-site.com | grep -o '<body>.*</body>'

# Bad result:
# <body><div id="root"></div></body>

# Good result:
# <body><main><h1>Your Content</h1>...</main></body>

If the body contains only a root div, your content is invisible to LLMs.

Frequently Asked Questions

Why doesn't ChatGPT cite my site?

If your site uses client-side rendering, the raw HTML is empty. LLM crawlers don't execute JavaScript, so they see nothing to cite.

Do SPAs hurt LLM indexing?

Yes. Pure SPAs with client-side rendering are invisible to most LLM crawlers. You need server-rendered or static content.

Can I use Lovable/Cursor and still rank in ChatGPT?

Yes, but you need to add a server-rendered layer for key pages. Deploy FAQ pages statically or use hybrid architecture.

Is Next.js bad for AEO?

No—Next.js supports SSR and SSG. The problem is when you use it in pure client-side mode.

How do I fix AI visibility quickly?

1) Add static/SSR for high-value pages. 2) Add JSON-LD schemas. 3) Ensure sitemap.xml is complete.

What is LLM retrievability?

Whether an LLM can find, fetch, and extract information from your content. Depends on training data, RAG access, and HTML parseability.

Take Action

Find out if LLMs can see your site:

Run Free AEO Audit Learn About Bullseye

What is AEO? – Complete guide
How to Rank on ChatGPT – Step-by-step
Bullseye AEO – Visibility tracking platform
LLM Lighthouse – Page-level analysis
Bull AI Home