Lovable SEO: What 189 Real Lovable Websites Get Right and Wrong

We audited 189 live Lovable websites for SEO. Newer builds ship real HTML with good titles, but none had a sitemap and almost none had a canonical tag or schema.

6

min read

Lovable logo above a Google search result for a Lovable site

Table of Contents

Lovable sites built since mid May 2026 are in good SEO shape. Older ones are invisible to most AI crawlers, and nearly all of them miss the same three basics.

Yes, mostly, if your Lovable site was built after mid May 2026. We audited 189 live Lovable websites and every one built since then ships its full content in the HTML, with a specific title, a meta description and a single h1. Sites built before that are client-side apps whose text only appears after JavaScript runs, which is a real problem for AI search.

Key findings

  • 38 of the 189 Lovable sites we audited, the ones built up to 10 May 2026, send crawlers an empty <div id="root">: the headline and text only appear after JavaScript runs.

  • Every Lovable site built from 17 May 2026 onward ships its content in the HTML, with a median of 603 words and its h1 already in place.

  • 0 of 189 Lovable sites had a sitemap.xml, only 3 had a canonical tag and 1 had structured data.

  • The page basics are solid: all 189 sites have a title and a meta description, and 183 of 189 have exactly one h1.

  • Lovable serves the same HTML to Googlebot, GPTBot and Chrome, and republishing an older site did not change its HTML in our test.

Every gap on this list can be fixed with a few lines in the prompt.

How we audited 189 Lovable sites

Every site in our library of Lovable examples was built by Lovable from a single prompt and published to a lovable.app address. For each one we fetched the raw HTML three times, as Googlebot, as Chrome and as OpenAI's GPTBot, to see what a crawler gets before any JavaScript runs. Then we loaded each page in a real browser and read it again after rendering. We also requested robots.txt and sitemap.xml for every site.

The audit ran on 13 October 2026. The two example sites shown below were fetched again on 16 October 2026 with a plain HTTP request using the GPTBot user agent, and their results had not changed.

Does Lovable render HTML for crawlers? Only since May 2026

This was the clearest result in the whole audit. Every site we built up to 10 May 2026 is a classic single page app: the HTML is an empty <div id="root"> and the text, headings and links only appear after JavaScript runs. Here is one of them, a yoga studio site:

Raw HTML of an older Lovable yoga studio site showing an empty root div, above the same page in a browser with the headline Find Your Stillness

Top: the body of the HTML that GPTBot receives from Solis Studio, an older build. Bottom: the same page in a browser after JavaScript has run. The headline "Find Your Stillness" appears nowhere in the HTML. Built from our original yoga studio prompt.

Every site built from 17 May 2026 onward ships the finished page, with a median of 603 words and its h1 right there in the HTML. A newer yoga studio build shows the difference:

Raw HTML of a newer Lovable yoga studio site with the h1 Move. Breathe. Transform. highlighted in green, above the same page in a browser

Root & Rise, a newer build of the same kind of site. The raw HTML already holds the h1 (green) and the h2 headings, 819 words in total before any JavaScript runs. Built from our yoga studio prompt.

Stacked bar chart of Lovable sites by build month: every site built up to 10 May 2026 renders only with JavaScript, every site from 17 May ships its content in the HTML

The 156 sites in our library with a known build date, by month in 2026. Grey sites show their content only after JavaScript runs, orange sites include it in the HTML.

Our prompts did not change in a way that would explain this, so it is a change on Lovable's side.

Two details matter. Lovable serves the same HTML to Googlebot, GPTBot and a normal browser, so there is no special version for crawlers. And republishing an old site did not change anything in our test. The old build stays a client-side app until the project itself is rebuilt.

Google does run JavaScript, so a client-side Lovable site can still rank, although Google describes rendering as a second, later step. AI crawlers are the bigger risk: research by Vercel found that the major ones do not execute JavaScript at all. To ChatGPT search or Perplexity, an older Lovable site can be a blank page.

What Lovable gets right for SEO

The on-page basics were better than we expected, on old and new sites alike:

  • Titles and descriptions: all 189 sites have a title and a meta description in the HTML, and almost all are specific to the business. Only 5 still say the default Lovable App, all of them early builds.

  • Sensible lengths: the median title is 52 characters and the median description 134, so most fit in search results without being cut. 14 percent of titles run over 60 characters.

  • Social previews: every site has its own og:image, a screenshot Lovable generates on publish, plus an og:title.

  • Structure: every site declares its language, and 183 of 189 have exactly one h1 with a median of six h2 headings below it.

Speed is in good shape too. The median site transfers 1.15 MB and paints its main content in about a second, as we measured in what Lovable websites weigh.

What Lovable leaves out: sitemap, canonical tag, structured data

  • No sitemap. 0 of 189 sites had a sitemap.xml. Every request returned a 404.

  • No canonical tag. Only 3 sites declared a canonical URL, which matters as soon as the site is reachable on both lovable.app and your own domain.

  • No structured data. 1 site had JSON-LD. For a local business that means no LocalBusiness markup with your address, phone number and opening hours.

  • Patchy robots.txt. All of the older builds had one. Of the newer server-rendered builds, 41 percent return the app's HTML page instead of a robots.txt file.

Bar chart of SEO checks on 189 Lovable sites: title, meta description and og:image 100 percent, one h1 97 percent, robots.txt 67 percent, canonical 2 percent, structured data 1 percent, sitemap 0 percent

SEO checks across 189 live Lovable sites. The page level basics are nearly perfect, the site level files are almost always missing.

This is what a crawler gets when it asks our two yoga studio sites for those files:

Requests for sitemap.xml and robots.txt on two Lovable sites: three return 404, only the older site has a robots.txt file

The newer Root & Rise build answers both requests with a 404 and its own HTML page. The older Solis Studio build has a plain text robots.txt but no sitemap. Built from our yoga studio and original yoga studio prompts.

Most of these sites are single page websites, because that is what our Lovable prompts ask for. The median site has 19 links and 89 percent of them jump to a section on the same page. That is fine for a small business site, but one URL can only rank for one main topic. If you want to rank for each service you offer, ask Lovable for a separate page per service.

The other pattern is placeholder links. On 117 of 189 sites at least one link points to #, usually social icons or footer links the prompt never gave a real address for. They do not hurt rankings much, but they are dead ends for visitors and they look unfinished.

How to prompt Lovable for SEO

Everything above can be fixed in the prompt before the first build. Add a block like this at the end, with your own domain and details:

SEO requirements:
- Give every page a unique <title> under 60 characters and a meta description under 155 characters.
- Add public/robots.txt allowing all crawlers and pointing to the sitemap.
- Add public/sitemap.xml listing every page on https://www.example.com.
- Add a canonical link tag on every page using https://www.example.com.
- Add LocalBusiness JSON-LD with the business name, address, phone, opening hours and URL.
- Every link must point to a real URL. Remove social icons that have no address

SEO requirements:
- Give every page a unique <title> under 60 characters and a meta description under 155 characters.
- Add public/robots.txt allowing all crawlers and pointing to the sitemap.
- Add public/sitemap.xml listing every page on https://www.example.com.
- Add a canonical link tag on every page using https://www.example.com.
- Add LocalBusiness JSON-LD with the business name, address, phone, opening hours and URL.
- Every link must point to a real URL. Remove social icons that have no address

SEO requirements:
- Give every page a unique <title> under 60 characters and a meta description under 155 characters.
- Add public/robots.txt allowing all crawlers and pointing to the sitemap.
- Add public/sitemap.xml listing every page on https://www.example.com.
- Add a canonical link tag on every page using https://www.example.com.
- Add LocalBusiness JSON-LD with the business name, address, phone, opening hours and URL.
- Every link must point to a real URL. Remove social icons that have no address

If the site is older than mid May 2026, check whether it is a client-side app first. If it is, ask Lovable to rebuild it on the current stack rather than just republishing it. For the rest of the prompt, our guide on how to write an AI website prompt covers structure and copy.

How to check your Lovable site's SEO in two minutes

  • View the page source (not the inspector) and search for your main headline. If it is not there, crawlers that skip JavaScript see an empty page.

  • Open /sitemap.xml and /robots.txt on your domain. A 404 or your homepage instead of a plain text file means they are missing.

  • Search the source for canonical and ld+json. If neither appears, add the lines from the block above.

What this audit does not cover

We measured technical SEO on published lovable.app sites, not rankings, and every site here was built from our own prompts, which ask for single page sites. A multi page Lovable project, or one connected to a custom domain, may behave differently. We tested republishing on one older site only. Other results on these same sites, from accessibility to how Lovable handles your design system, are in our other studies.

FAQ

Is Lovable good for SEO?

For sites built since mid May 2026, mostly yes. In our audit of 189 Lovable sites, newer builds ship full HTML with a specific title, a meta description and one h1. What every site missed was a sitemap, a canonical tag and structured data.

Is Lovable good for SEO?

For sites built since mid May 2026, mostly yes. In our audit of 189 Lovable sites, newer builds ship full HTML with a specific title, a meta description and one h1. What every site missed was a sitemap, a canonical tag and structured data.

Can Google index a Lovable website?

Yes. Google runs JavaScript, so even older client-side Lovable sites can be indexed. Sites built since mid May 2026 are easier, because their content is already in the HTML.

Can Google index a Lovable website?

Yes. Google runs JavaScript, so even older client-side Lovable sites can be indexed. Sites built since mid May 2026 are easier, because their content is already in the HTML.

Can ChatGPT and other AI search engines read Lovable sites?

Newer ones, yes, because the text is in the HTML. Older Lovable sites that only render with JavaScript can look empty to AI crawlers, which generally do not run JavaScript.

Can ChatGPT and other AI search engines read Lovable sites?

Newer ones, yes, because the text is in the HTML. Older Lovable sites that only render with JavaScript can look empty to AI crawlers, which generally do not run JavaScript.

Does Lovable create a sitemap?

Not by default. None of the 189 sites we audited had a sitemap.xml. Ask for one in the prompt, together with robots.txt and a canonical tag.

Does Lovable create a sitemap?

Not by default. None of the 189 sites we audited had a sitemap.xml. Ask for one in the prompt, together with robots.txt and a canonical tag.

Get new AI website prompts every week — straight to your inbox.

Every Friday we drop a fresh batch of AI website prompts and tips to help you build better websites faster.

Join 1,000+ AI enthusiasts

liquid metallic gradient

Get new AI website prompts every week — straight to your inbox.

Every Friday we drop a fresh batch of AI website prompts and tips to help you build better websites faster.

Join 1,000+ AI enthusiasts

liquid metallic gradient

Get new AI website prompts every week — straight to your inbox.

Every Friday we drop a fresh batch of AI website prompts and tips to help you build better websites faster.

Join 1,000+ AI enthusiasts