Why your Lovable site is missing from Google or ChatGPT
Published · Last updated
As of November 2026, the answer depends on when your Lovable project was made. Projects made on or after May 13, 2026 use TanStack Start and render every page on the server, so crawlers get the full page. Older projects use React and Vite: the server sends an almost empty page, and Lovable pre-renders it only for verified search engines and social preview bots. Lovable's own documentation says other tools see very little, and AI crawlers such as GPTBot and ClaudeBot do not run JavaScript.
Check which kind of project you have
- In Lovable, open Project settings, then General, then Project actions. If you see Migrate to TanStack Start, your project is on the older stack. If the option is not there, the project is probably on TanStack Start already.
- Open your published site, choose View Source (not Inspect) and look for your headline and main paragraph. If they are not in that raw HTML, a crawler that does not run JavaScript cannot see them.
- Do this for more than the home page. Pages that load data after the page opens can be empty even on a server-rendered project.
Why you can be in Google but not in ChatGPT
On an older project, Google gets a pre-rendered copy, so your pages can rank in Google. AI crawlers are not on that list, so they can get the empty page. The same page can then show up in Google results and still be missing from the answers in ChatGPT, Claude and Perplexity.
Three ways to fix it
- Upgrade to TanStack Start. Lovable can upgrade an older project in place; Lovable's upgrade guide says it usually costs 10 to 35 credits, and your live site does not change until you publish. Test every page after the upgrade, because code that only works in a browser can break server rendering.
- Use a prerender service, and make sure it serves the rendered page to every AI crawler, not only to Googlebot. Some setups serve it to a short list of bots and leave the others on the empty page.
- Make the pages you want AI assistants to cite into server-rendered or static pages, and leave the app itself as it is.
A readable page is only the first step
A page that crawlers can read must also be found and worth citing. Each page needs its own title, description and canonical link, a place in your sitemap and a link from another page on your site. Search engines must know about it (see the indexing guide). AI assistants cite pages that answer a buyer's question directly in the first paragraph.
What this does not tell you
Which of your pages each AI crawler can read today, whether your pre-render covers every AI crawler, and what to fix first on your site.