SEO Checklist for AI-Built Websites Before Launch

An AI-built website can rank on Google when crawlers receive the real page content, each page can be indexed under one URL on your own domain, and the pages answer what customers search for. Before launch, check what the raw HTML contains, confirm the robots.txt, sitemap, noindex, and canonical settings, give every page its own title and description, link the pages to each other, and connect Google Search Console. The checklist below applies to any builder. The numbers in it come from our crawl of 417 public template sites from Lovable and Base44.

Key takeaways

• Builders handle different parts of SEO for you. In our crawl, 154 of 189 Base44 website templates served an XML sitemap, compared with 9 of 228 Lovable templates.
• In most templates from both builders, the visible page wasn't in the raw HTML. Both builders say they serve crawlers a rendered version, so confirm what Google receives with URL Inspection.
• A template can carry settings meant for someone else's site. Six Base44 template homepages were marked noindex, and one had a canonical tag pointing to a different site.
• Google allows AI-written content. Pages produced in bulk without adding value can break its spam policy on scaled content abuse.
• Put the site on a custom domain before you launch. Base44 says its SEO setup is designed for custom domains.

Define the searches each page should answer

Settings get a page considered for search. What the page says decides which searches it can rank for, so start there. Write down one search for each page you plan to publish, in the words a customer would type: "emergency plumber in Tucson" for a service page, "kitchen remodel cost" for a pricing guide, or your business name for the homepage.

If two pages share the same search, merge them or give one a different job. If an important search has no page, plan one. Many templates start as a single page: 44 of the 154 Base44 sitemaps we read listed only the homepage. A one-page site gives Google one URL for every service you offer. Separate pages for each main service or location, each with its own URL and content, are easier to match to specific searches.

What two builders' templates ship with

To see the defaults a new project inherits, we fetched every site linked from Lovable's template gallery and from the websites section of Base44's template gallery on October 17, 2026. We requested each homepage, robots.txt, and sitemap.xml once, the way a simple crawler or link checker would, and recorded the raw HTML before any JavaScript ran.

The other 152 Base44 sites served a hidden block containing the page heading, the description, and links to the app's other pages. Base44's search visibility documentation describes this block as the fallback crawlers receive when a page hasn't been rendered yet. It says Base44 otherwise serves crawlers a rendered version with the real headings and content. Lovable's SEO documentation says projects created since May 13, 2026 render on the server, and older projects are pre-rendered for verified crawlers on Lovable hosting.

Read the table with that in mind. Our client wasn't a verified crawler, so Googlebot may have received more than we did. Templates are starting points, and a finished site should differ from them. Every Base44 site we checked was on a free base44.app address, and Base44 says those addresses "are not intended for long-term, production SEO". The table shows which settings a new project starts with, not what Google indexes.

URLs and rendering

Look at your page the way a basic crawler does. Open the browser's view-source on your homepage, or run curl -s https://yourdomain.com/, and search for your main heading and a sentence of body text. If they aren't there, the content is built in the browser or served only to crawlers the builder recognizes.

That can still work. Use Search Console's URL Inspection live test on your custom domain and open the rendered HTML. If your headings, text, and links are there, Google can read the page. If they're missing, fix rendering before working on anything else.

Rendering can change when a site moves. Lovable pre-renders older projects only on its own hosting, and Base44 describes its crawler rendering as a feature of its platform. Before moving an exported project to other hosting, test how the new host delivers the pages. Our prototype to production guide covers the rest of that move.

Two structural checks apply to every builder:

• Each page should have its own URL that loads directly, without first opening the homepage.
• Navigation should use real links (<a href="/services">) that crawlers can follow, rather than buttons that change the view with a script.

Indexability and canonicals

These settings decide whether Google keeps a page. Check them on the custom domain after you connect it, because some values change when the domain changes.

robots.txt tells crawlers which paths they may visit. Open /robots.txt and confirm it doesn't disallow pages you want found. If it has a Sitemap: line, Google's robots.txt documentation says it "must be a fully qualified URL, including the protocol and host". One Base44 template used Sitemap: /sitemap.xml. robots.txt asks crawlers to stay away from a path, but it doesn't stop anyone from opening the page. Admin and login pages need a login to protect them.

A noindex rule removes a page from search. It can be set in a <meta name="robots"> tag or in an X-Robots-Tag response header, according to Google's guide to blocking indexing, so check both:

curl -sI https://yourdomain.com/ | grep -i x-robots-tag
curl -s https://yourdomain.com/ | grep -io '<meta name="robots"[^>]*>'

Six of the 189 Base44 template homepages sent noindex, follow in both places. Base44 adds noindex to any page whose Search visibility toggle is off, so if you started from a template, check the homepage and every page you plan to publish.

A canonical tag names the URL Google should keep when the same content is reachable at more than one address. Google recommends a canonical tag on the canonical page itself, pointing to its own URL, and warns against different signals for the same page, such as one URL in the sitemap and another in the tag. Every page should point to its own address on your custom domain. One Base44 template, wordle-world-template.base44.app, had a canonical pointing to wordle-world.base44.app, a different site. A project copied from it would tell Google that another site is the real page until someone changes the tag.

The XML sitemap lists the pages you want indexed. Google's sitemap guidance asks for "fully-qualified, absolute URLs". One Base44 template listed paths such as / and /book without the domain. Open /sitemap.xml and confirm each entry is a full URL on your custom domain and that private pages aren't listed.

Titles and on-page structure

Each page needs its own title and meta description that name what the page offers. Generated sites tend to set them once for the whole app, or leave the builder's default. In our crawl, 32 Lovable templates had a default or missing title, and 18 Base44 templates had no meta description.

Google's guidance on generative AI content asks site owners to fact-check AI-generated content before publishing, and it names titles, meta descriptions, structured data, and image alt text as part of that review. Read each one as a customer would see it in search results.

On the page itself:

• Use one H1 that matches the page's search, and H2 headings that describe each section.
• Put the important text in HTML, rather than inside images or content that only appears after a click.
• Write alt text that describes each meaningful image.
• Set a share title, description, and image for each page, so links posted in messages show the right preview.

Base44 added WebSite and Organization structured data to 87 of the templates. After you connect your domain, check that the name and URL in that markup are your business name and your domain. Add other types, such as LocalBusiness or Product, only when they describe what the page visibly shows.

Internal links

Crawlers discover pages by following links and reading sitemaps. Link every important page from your navigation or from related pages, and use anchor text that names the destination, such as "water heater repair" instead of "learn more". A page that nothing links to may be found through the sitemap, but it gives Google little context about where it fits.

Keep admin, login, and account pages out of public navigation and the sitemap. One Base44 template's crawler summary linked to /admin, the same path its robots.txt asked crawlers to skip.

Local business relevance

Google's local ranking help says local results are "mainly based on relevance, distance, and popularity", and it points businesses to a verified, complete Business Profile with reviews. Your website supports that profile when it states the same business name, address, phone number, and hours, and when each service or location page explains what you do there.

Avoid location pages that repeat the same text with a different town name. Google's spam policies list pages "targeted at specific regions or cities that funnel users to one page" as an example of doorway abuse. Write a location page only where you have something specific to say, such as the neighborhoods you serve, local projects, or how scheduling works in that area.

Image and mobile performance

Google's Core Web Vitals guidance sets good targets of the main content loading within 2.5 seconds, responses to taps within 200 milliseconds, and a layout shift score below 0.1. Test your main pages in PageSpeed Insights on mobile.

Common causes include large uncompressed images, animation libraries loaded on every page, and images without a set width and height, which make the layout jump as they load. Compress images, give each one dimensions, and load images below the first screen lazily.

Analytics and Search Console

Set these up on launch day, so you have data from the start:

1. Add your custom domain to Google Search Console and verify it. Base44 and Lovable both offer connectors that verify the site and submit the sitemap, and you can also do it directly.
2. Submit the sitemap URL.
3. Install analytics and record form submissions and calls as events. Keep names, email addresses, and message text out of analytics.
4. Inspect your most important URLs and request indexing for them.

Monitoring after launch

New sites take time to be indexed. Base44 says it can take from a few days to a few weeks. Use this schedule for the first month and after any large change:

• First week: run URL Inspection on the homepage and main service pages to confirm Google can fetch and render them.
• Weeks two to four: open the Pages report and read the reason Google gives for each URL that isn't indexed.
• Monthly: check the Performance report for the searches people use to find you, and compare them with the list you started with.
• After republishing or changing settings: recheck noindex, canonical, and sitemap. Base44 notes that if you turn its main SEO setting off and back on, any page that was switched off stays hidden from search until you switch it on again.

The pre-launch checklist

Work through this on the custom domain, one page at a time for the pages that matter most.

Before you publish:

• Each page has one search it should answer, and no two pages compete for the same one.
• The homepage and main pages each have their own URL that loads directly.
• Titles, meta descriptions, share previews, and alt text are unique and reviewed by a person.
• Each page has one H1 that matches its search.
• Location and service pages contain details specific to that location or service.

At launch:

• The site is published, public, and on your custom domain.
• robots.txt doesn't block public pages, and any Sitemap: line is a full URL.
• No public page has noindex in a meta tag or an X-Robots-Tag header.
• Each canonical tag points to the page's own URL on your domain.
• The sitemap lists only public pages, as full URLs on your domain.
• Structured data names your business and your domain.
• URL Inspection's rendered HTML shows your headings, text, and links.
• Search Console is verified, the sitemap is submitted, and analytics records form submissions.

In the first month:

• Main pages pass Core Web Vitals on mobile, or you know which images or scripts to fix.
• The Pages report has no unexplained "Excluded" reasons for important URLs.
• Search Console shows impressions for the searches you planned.

When developer help is useful

Most of this checklist is settings and writing, which you can do yourself. A developer is worth bringing in when URL Inspection's rendered HTML is missing content, when you need to move off the builder's hosting without losing rendering, when structured data has to come from your data for each page, or when speed problems trace back to how the app is built. If you built on Lovable, the Lovable SEO guide covers its settings in detail. Our page on finishing AI-built projects explains how we review a site built with any builder, and the guide to hiring a developer to finish an AI website covers what to ask before you hire.

Frequently asked questions

Can Google index AI-built websites?

Yes, when crawlers can read the content and the pages are public, published, and indexable. Lovable and Base44 both serve crawlers rendered pages on their own hosting. Check your own pages with URL Inspection's live test, which shows the HTML Google rendered and any reason a page wasn't indexed.

Does AI-written copy rank?

Google's guidance allows AI-generated content and judges it by the same standards as any other page. It warns that generating many pages without adding value for users may violate its policy on scaled content abuse, and it asks owners to fact-check AI output, including titles and descriptions. Copy that answers a specific search with accurate details is what can rank, whoever drafted it.

Do I need a custom domain for SEO?

You can be indexed on a builder's free address, but a custom domain is the safer choice for a business. Base44 says its SEO setup is designed for custom domains, and Lovable requires a paid plan to connect one. Moving later means redirecting every indexed URL to the new domain and waiting for Google to process the change.