Is your Lovable or Bolt site blank to AI crawlers? A 2-minute check

If you built your site with Lovable, Bolt, Replit, Emergent or a similar tool, there is a good chance it is a single-page app. Your browser downloads a nearly empty page, then JavaScript fills in the text. You never notice, because your browser runs the JavaScript.

Many tools don't. Link previews on WhatsApp, Slack and LinkedIn read the raw page. So do many AI crawlers. Google does run JavaScript, but later and not always fully. If your key text only appears after JavaScript, some of the tools that decide whether people find you see a blank page.

In one of our own apps, the homepage sent 10 words in its raw HTML. After JavaScript, it had 709. (Full case study.)

The 2-minute check

Open your homepage in Chrome or Safari.
View the page source: right-click and choose View Page Source, or put view-source: in front of the address.
Press Ctrl+F (Cmd+F on a Mac) and search for a sentence from your headline.

Found it? Good. Your text is in the page the server sends.

Not found, and the source is mostly a <div id="root"> and script tags? Your text only appears after JavaScript runs. Repeat the check on two or three other important pages, such as pricing or a product page.

A stricter check, if you are comfortable with a terminal

This counts the words in the raw HTML, ignoring scripts and styles:

curl -s https://yoursite.com | python3 -c 'import sys,re;h=sys.stdin.read();h=re.sub(r"(?is)<(script|style|noscript|template)[^>]*>.*?</\1>"," ",h);print(len(re.sub(r"<[^>]+>"," ",h).split()),"words")'

A homepage that shows hundreds of words in your browser but single digits here has the problem.

What to ask your builder

Paste this, adjusted to your pages:

Please make the public pages send their main content in the HTML the server returns, before JavaScript runs: the heading, the main text, the page title, the meta description, the canonical tag and the preview (Open Graph) tags. Use pre-rendering or static generation at build time. Generate the list of pages from the same data the app uses, not by hand. Keep sign-in, account and other private pages exactly as they are: not pre-rendered, not indexed and not in the sitemap. The app must still work normally in the browser, with no double render. After deploying, show me the word count of the raw HTML for the homepage and two other pages.

The last line matters. Ask for proof, then check it yourself with the 2-minute test.

What this won't do

Fixing this makes your site readable. It does not guarantee rankings or mentions in ChatGPT, and anyone who promises that is guessing. But a page that tools can't read can't be found or quoted at all.

Is your site readable by search engines and AI tools?

Run the same checks on your site inside Claude, Cursor, VS Code or Gemini CLI. See the evidence, and get the exact fix for your builder.

Check my site free

Haigreeva Kondury

Founder of Visibility Readiness, based in Bengaluru. Background in risk and compliance, where every claim needs evidence. Builds with AI tools and doesn't write code. More about us