- Lovable checks projects under “More → SEO & AI search” for indexing, robots.txt, sitemap, metadata, structured data, content structure and readability for AI crawlers; the scan is free on all plans, while fixes use credits.
- New Lovable apps (since May 13, 2026) render on the server; older React and Vite apps serve prerendered HTML to verified crawlers only.
- Lovable no longer checks load time, accessibility or mobile-friendliness itself; it recommends PageSpeed Insights instead.
- You don't need an llms.txt, neither for Lovable nor for Google Search: according to Google, optimizing for AI search is still SEO.
16:00Lovable lets you build websites and apps quickly, without anyone writing every line of code. As a result, far more new projects are going live, and the competition for visibility is getting tougher. It used to be enough to optimize for Google and Bing. Today, projects also need to be readable for AI-powered answer engines such as ChatGPT, Perplexity or Google's AI Mode.
For this, Lovable offers its SEO & AI search section, which scans a project and can fix many issues directly. In this article, we show what the scan checks, how to read the results and where the decisions are still up to you.
SEO and AI search share the same foundation
To be found on the web, a project needs two things:
- Search engine optimization (SEO): your project ranks high on Google or Bing.
- Answer engine optimization (AEO): AI systems like ChatGPT, Perplexity, Claude or Gemini understand your project, use it as a source and recommend it.
Both share the same fundamentals: clean HTML, precise metadata, a clear content structure, fast load times and links from other websites. Google confirms this in its July 2026 guidance: optimizing for AI Overviews and AI Mode is still SEO. Lovable handles much of the technical work; the strategic decisions remain yours.
Check and improve your project in four steps
You'll find it in the editor's project toolbar under More → SEO & AI search:
- Scan the project: Click Scan project, or Scan again on later runs.
- Review the issues: The findings that pose the greatest risk to your visibility appear at the top of the report.
- Let Lovable fix them: Try to fix has the Lovable agent apply the fix for a single finding; Try to fix all fixes all of them at once. Use Ignore to hide findings that aren't relevant to your project.
- Publish and scan again: Checks against the live site, such as indexing and Google Search Console, only run after publishing. Lovable doesn't rescan automatically after you publish; you have to start the scan yourself.
The scan checks nine areas of your project
Lovable combines checks of the code, the preview and the published site:
| Area | What is checked | Common problems |
|---|---|---|
| Page basics | Basic page structure | missing viewport tag or missing langattribute |
| Indexing | Search engine access | unreachable homepage, site-wide noindex, Search Console not set up |
| robots.txt | Crawler rules | missing file, blocking rules, missing reference to the sitemap |
| Sitemap | Completeness and validity | missing sitemap, invalid XML, placeholder routes, relative URLs, wrong host |
| Metadata | Titles, descriptions, canonical | duplicate titles, placeholder text, weak descriptions, incorrect canonical URLs |
| Open Graph | Link previews when sharing | missing details or Lovable branding in shared links |
| Structured data | JSON-LD | invalid or unsuitable markup, missing markup for rich results |
| Content structure | Headings, images, links | missing H1, incorrect heading levels, missing alt text, vague link text |
| AI readability | Output for AI crawlers | confirms that AI crawlers receive a clean Markdown version |
Lovable no longer checks load time, accessibility and mobile-friendliness itself. For this, Lovable recommends Google's PageSpeed Insights. Check load time and mobile usability after major changes. Technical problems can prevent visitors from using your offer or getting in touch.
The icons show you where action is needed
After each scan, Lovable groups the results by severity:
- Green check mark: passed, no action needed.
- Blue light bulb: low impact, worth fixing.
- Yellow warning triangle: medium impact, fix soon.
- Red X: high impact. Examples include a site-wide
noindex, an unreachable homepage or arobots.txtthat blocks crawlers. Errors like these can remove your website from search results entirely.
Lovable also marks findings as Fixed once the agent has resolved them; the next scan confirms this. A label also shows whether the scan reflects the current state (Up to date) or whether you've changed something since (Out of date).
Set up Search Console and Semrush right in the chat
Connect Google Search Console in a few steps
If the Search Console connector is enabled in your workspace, Lovable checks whether your site is connected to Google and verified, and whether a sitemap has been submitted. If anything is missing, click Try to fix: Lovable connects Search Console, verifies the site and submits the sitemap straight from the chat.
Publish your latest changes first, since Google fetches the sitemap from the live site. We cover this in detail in our article on connecting Search Console and Lovable.
Research search terms and competitors with Semrush data
For strategic questions, Lovable taps into live data from Semrush. Until November 30, 2026, this comes at no extra cost, and you don't need your own Semrush account. You can use the suggested prompts or ask your own, for example:
- “Which search terms should a website like mine target next?”
- “Which competitors rank for similar products?”
- “How can I improve my landing page copy for search?”
The answers are a good starting point. But only by checking them against your own data will you see which search terms actually drive inquiries and revenue.
A custom domain consolidates your visibility
A professional presence needs its own domain. Under Get a custom domain, you can buy a domain or connect one you already own; buying one requires a paid plan. Once you set a primary domain, all other connected domains redirect to it.
After switching to your custom domain, you'll need to re-verify Search Console and rerun the scan so the sitemap and settings match the new address.
Check what crawlers actually see of your project
What matters for visibility is whether search engines and AI crawlers receive complete HTML. Many AI crawlers don't execute JavaScript. Lovable handles this differently depending on how old the project is:
- Newer apps (created on or after May 13, 2026) use TanStack Start with server-side rendering. Every request returns fully rendered HTML.
- Older React and Vite apps use on-demand prerendering. Only verified crawlers receive the prerendered HTML, such as those from Google, Bing, social media preview services and AI systems like ChatGPT, Perplexity, Claude and Gemini. All other requests get the standard JavaScript app.
For older projects, that means crawlers not on this list, and many SEO tools, see a blank page. For important projects, we therefore recommend considering a migration to TanStack Start. It costs credits, and the live site only changes once you publish again. Our article on the JavaScript problem of AI crawlers explains why JavaScript is an issue for AI crawlers.
You don't need an llms.txt
For a long time, an llms.txt file was considered essential for AI search. Lovable no longer flags a missing one as an issue; an existing file will still be served. Google's guidance on AI search makes clear that Google Search doesn't use such files.
Visibility in AI answers depends on the same things as SEO: content that answers questions directly and with evidence, your own data and experience, and mentions on other websites. Learn more in our article on Generative Engine Optimization.
Fundamentals to know before you launch
- Costs: The scan is free on all plans. Only fixes made by the agent use regular credits.
- Unpublished projects: You can scan before publishing for the first time. Checks against the live site, such as indexing, only run afterward.
- Indexability: Only publicly published apps can appear in search engines.
Conclusion: Lovable handles the tech; visibility comes from content
The SEO & AI search section takes care of many technical fundamentals for you: indexing, sitemap, metadata and the Search Console connection. That's a solid foundation, but it isn't visibility yet. As Lovable produces more and more projects that are equally clean technically, it's positioning, substantive content and mentions on other websites that determine who gets found and recommended.
If you want to set up a Lovable project to perform in search and AI answers, we can help with SEO and GEO.
