Blog / Technical Guides

From Lovable Prototype to Real Business Website: A Launch Checklist

By Keyona Meeks, Founder, ReRev Labs · Oct 31, 2026 · 9 min read

The short answer: A Lovable prototype becomes a real business site when eight things are true: it lives on your own domain, one address is the main one, every page has its own title, forms reach a human, search engines can find the pages, analytics is recording, you can see how Google sees you, and links shared on social sites show a preview. Each has a check you can run in a few minutes. This checklist is my own, built from the checks in our guides and from our own site. It does not replace the vendors' documentation, and I could not read several of them. We sell build work, so weigh this with that in mind.

By the end of this post, you'll be able to:

  • Run eight launch checks on a Lovable site.
  • Tell a cosmetic problem from one that costs you leads.
  • Know what you could not verify without the vendor's page.
  • Decide what to fix first.

I sell build work. This is the overview. Each check has a longer guide, and I say what I could and could not confirm for each. Nothing here is a guarantee about search rankings.

Is the site on your own domain?

A prototype often lives at a builder's address. A business site needs your own name. Our post on deploying a Lovable website to your own domain describes three routes: Lovable's built-in option, a GitHub and Netlify route, and your own hosting. For how a domain points at a host, see the DNS crash course for founders and clients.

Check: open your site from a phone on mobile data, not your office wifi. Does your own address load, with a padlock?

Before you move DNS, write down every existing record. Email often depends on them.

If the domain fights back: custom domain not working covers pending, SSL and www errors, 404 on refresh is the single-page-app redirect fix, and there are per-host walkthroughs for Vercel and Cloudflare Pages.

Is one address the main one?

www.yourname.com and yourname.com are different addresses. Pick one, and send the other to it with a permanent redirect. Google's guidance, as I read it in excerpts, is to pick one address as canonical. Our guide on choosing www or no www covers the redirect and four curl commands to check it.

Check: run curl against all four variants (http and https, with and without www). Each should end at your chosen address, and only that address should return 200. On our own site on October 31, 2026, three variants returned a 301 and the main address returned 200.

The fix: pick www or no www and redirect the other.

Does every page have its own title?

A single-page app ships one HTML file with one title. Without a fix, every page can start with the same title. Our guide on the same-title problem explains a browser-side fix and a build-time prerender.

Check: fetch three pages with curl and read the raw <title>. They should differ. On our site, three pages returned three titles on October 31, 2026. I could not read Google's title-link page, so I make no claim about how Google picks titles.

The fix: when every page shows the same title.

Does the contact form reach a person?

A form on a page cannot send email by itself. Something must receive it. The failure is silent: the visitor sees a thank-you, and you see nothing. Our guide on contact form notification gives a small Google Apps Script that emails you once per submission, tested with a mock, not in Google. If you use an embedded MailerLite form, see the MailerLite and Lovable trap.

Check: submit the live form yourself from a different device and email. Does it arrive, and how fast? Do this again after every site change.

Two builds: get notified without a paid tool, or send submissions straight to Google Sheets.

Can search engines find your pages?

Two plain files help: a sitemap and a robots.txt. Google's sitemap page says "All formats limit a single sitemap to 50MB (uncompressed) or 50,000 URLs" and "submitting a sitemap is merely a hint." Its robots.txt page says the file "is not a mechanism for keeping a web page out of Google." Both are excerpts. Our guide on sitemaps and robots.txt has a small tested script.

Check: fetch /sitemap.xml and /robots.txt with curl. You want the right content, not your home page.

The builds: add a sitemap and robots.txt, make the site AI-crawler friendly, and read what we learned fixing our own Lovable SEO.

Is analytics recording?

The tag can stop without warning. Our guide on Google Analytics 4 describes how our own site loads the tag after the visitor accepts, and gives a weekly formula that flags a zero or a sharp drop. I could not read Google's Analytics help pages, so I give no Analytics menu paths.

Check: once a week, compare yesterday's sessions with the average of the week before. A zero is a flag. Ask your attorney whether you need consent handling for your visitors.

The build: add Google Analytics 4 and check it is actually recording.

Can you see how Google sees you?

Google offers a tool, Search Console, for this. I could not read its help pages in this session, so I give no click path and make no claim about what it shows. Look up current steps on Google's own site.

Check: after you set it up, make sure the property matches your chosen address, with or without www, so you are not looking at the wrong one.

The build: connect Search Console and read the first report.

When someone pastes your link into a chat or a social post, the preview usually comes from tags in the page. The Open Graph site lists four required properties: og:title, og:type, og:image and og:url. It describes the image property as "An image URL which should represent your object within the graph" and the URL as "The canonical URL of your object that will be used as its permanent ID in the graph." Source: The Open Graph protocol, read as excerpts on October 31, 2026.

Check: fetch a page with curl and look for those tags. On October 31, 2026, one of our blog pages returned og:title, og:type, og:url, og:image and an og:description in the raw HTML. I did not test how any chat app or social site shows the preview, and previews may be cached, so a fix may not show at once.

The fix: when the link preview does not show.

What should you fix first?

My order, by what costs you leads:

  1. The form. A form that sends nothing loses people you paid to attract.
  2. The domain and the main address. A broken or duplicated address confuses everyone.
  3. Analytics. If you cannot see visits, you cannot judge anything else.
  4. Titles and the sitemap. They help people find you.
  5. Search Console and link previews. They help you see and present.

This order is a judgment. Change it if your site is mostly shared on social media and not found through search.

Who owns it after launch?

Name one person, and write down where everything lives: the domain account, the host, the form script, the analytics property. Check the list monthly. A site nobody owns drifts. For the idea of a one-page note that others can read, see the one-pager that gets things done.

Where does an LLM fit in this?

An LLM can help with the messy part: reading your site's pages and drafting titles and descriptions, or reading a list of links and flagging the ones that use the wrong address. A person should check each one, because an LLM can write a plausible title that says something untrue about the page. The checks in this post are plain commands and rules, and they do the same thing every time. My stance: AI writes the code, plain automation runs it, and an LLM shows up only where the input is messy and a person checks the result. For more on what to automate, see what to automate first.

Where this has limits

  • This is my checklist. It is not a standard and it is not exhaustive.
  • I could not read Lovable's documentation or several Google pages. That includes Search Console, Analytics and title-link help. I give no click paths for them.
  • The Google and Open Graph lines are excerpts. Check the live pages.
  • The test results for our own site are from October 31, 2026.
  • Nothing here is a guarantee about rankings or leads.
  • Privacy and consent rules vary. Ask your attorney.
  • We sell build work. Read this with that in mind.
  • No ReRev prices and no promises about results.

If you want help taking a Lovable site live, or want to claim a free rebuild of a tool you pay for, tell us what you use. Do not send passwords or login details.

Frequently asked questions

What should I check before launching a Lovable website?

The domain, the main address, page titles, the contact form, the sitemap and robots.txt, analytics, search visibility and link previews.

What is the most important check?

The form. If it sends nothing, you lose the people the site was built to reach.

Do I need a sitemap?

It can help. Google calls submitting one "merely a hint."

How do I know my analytics is working?

Check the tag is on the page and look for your own visit. Then compare each week's numbers so a silent stop shows up.

Will these checks improve my rankings?

I cannot promise that. They remove common problems.

Who should own the site after launch?

One named person, with a written list of where everything lives.