Website launch checklist: 23 things to check before you share your site
By Roshan, founder of SearchMaxx - - 5 min read
A site can look finished and still be broken for search engines, link previews and AI tools. The usual causes are small: a staging noindex left in, an app that returns 200 for every URL, or a share image that 404s.
This is the list we run on every site before we start work. Each item says why it matters and how to check it with curl. Replace paidwick.com with your domain.
To run all 23 checks at once, use the free website launch checklist. It takes about 10 seconds.
Loads and redirects
1. The homepage loads over HTTPS. Every other check depends on it.
curl -sI https://paidwick.com/ | head -1
You want HTTP/2 200.
2. HTTP redirects to HTTPS. Old links and typed addresses still use http://. They should get a 301 to the secure page.
curl -sI http://paidwick.com/ | grep -iE '^(HTTP|location)'
You want 301 and a location: https://... line. If you see 200, turn on your host's "always use HTTPS" setting.
3. www and the bare domain lead to one site. Pick one and 301 the other to it. Two live copies split links, and Google has to guess which one is canonical.
curl -sI https://www.paidwick.com/ | grep -iE '^(HTTP|location)'
If the domain does not resolve at all, add a DNS record and redirect it. People do type www.
4. Missing pages return a 404. Request a URL that cannot exist:
curl -s -o /dev/null -w '%{http_code}\n' https://paidwick.com/this-page-does-not-exist-123
You want 404. Single-page apps often return 200 with the app shell for every path. Google calls this a soft 404. The router in the browser can show a "not found" screen, but the server or host still has to send the 404 status.
5. The homepage responds quickly. Time to first byte is how long the server takes to start answering.
curl -s -o /dev/null -w '%{time_starttransfer}\n' https://paidwick.com/
web.dev suggests 0.8 seconds or less as a rough guide. Our checklist flags anything over 1 second, measured from our server.
Search basics
6. A title tag. It becomes the title link in Google and is often the name AI answers use for the page. Name the product and what it does.
7. A meta description. Google often uses it for the snippet under the title. Without one, it picks text from the page. Aim for 70 to 160 characters.
curl -s https://paidwick.com/ | grep -ioE '<title>[^<]*</title>|<meta name="description"[^>]*>'
8. A canonical tag that points to the page itself. Copied templates often point every page at the homepage, or at a staging domain.
<link rel="canonical" href="https://paidwick.com/">
9. No noindex. A noindex rule keeps the page out of Google. It can be a meta tag or an X-Robots-Tag header, so check both:
curl -sI https://paidwick.com/ | grep -i x-robots-tag
curl -s https://paidwick.com/ | grep -io '<meta name="robots"[^>]*>'
10. One h1. It should be in the HTML the server sends, not only added by JavaScript. Crawlers that do not run scripts never see it.
11. A language on the html tag. <html lang="en"> tells browsers and screen readers what language the page is in.
12. A mobile viewport tag. Google indexes the mobile version of your site. Without this tag, phones render a shrunken desktop page.
<meta name="viewport" content="width=device-width, initial-scale=1">
13. A robots.txt that allows search. A missing file is fine for crawlers. A leftover Disallow: / from staging blocks the whole site.
curl -s https://paidwick.com/robots.txt
If you see your HTML page instead of a short text file, your app is answering every path. Fix that first. Our guide to which AI crawlers to allow has a robots.txt you can copy.
14. A sitemap, listed in robots.txt. A sitemap lists the pages you want found. Add one line to robots.txt so every crawler finds it:
Sitemap: https://paidwick.com/sitemap.xml
15. Structured data that parses. JSON-LD names your company and product in a format machines read directly. A missing comma breaks the whole block, so paste your homepage into Google's Rich Results Test after every change. Our post on schema markup for SaaS covers which types to use.
Sharing
16. Open Graph title and description. Slack, LinkedIn, iMessage and most chat apps build link previews from the Open Graph tags.
17. A share image that loads. The tag is easy to add and easy to break. Check that the URL returns an image:
curl -sI https://paidwick.com/og.png | grep -iE '^(HTTP|content-type)'
Use 1200x630 and a full https:// URL. Some platforms ignore relative paths.
18. An X card tag. X needs twitter:card to show a large image. It reads the Open Graph tags for the rest.
<meta name="twitter:card" content="summary_large_image">
19. A favicon. It shows in tabs, bookmarks and next to your result in Google. Add an apple-touch-icon too for iPhone home screens.
AI search
20. An llms.txt. A proposed convention from llmstxt.org: a Markdown file at /llms.txt with your name as an h1, a short summary, and links to your key pages. Google says its search ignores the file, and no AI search company says its crawler reads it. It helps coding tools and agents find your key pages, and it takes an hour to write. Our llms.txt guide has the evidence and a template.
curl -s https://paidwick.com/llms.txt | head -3
The first line should be a # and a space, then your name. If you get HTML, your app's catch-all route is answering instead.
21. An llms-full.txt (optional). The full text of your docs or key pages in one Markdown file. It is newer and less common than llms.txt, so treat it as a nice extra.
Trust
22. A privacy policy link. Buyers look for one, and so do app stores and ad platforms. Put it in the footer.
23. A contact or about link. People check who runs a site before they buy. An email address in the footer counts.
What this list does not cover
Passing all 23 means nothing is stopping search engines, AI tools and link previews from reading your site. It does not get you ranked or named in AI answers. That depends on what your pages say and who links to them.
For the AI side, run the free AI site check. It tests whether ChatGPT, Claude and Perplexity's crawlers are allowed in and get the same page as people.
Want this done for your site?
Every day we make one change to your website that helps AI engines name you, and a person checks each one. Email me your website and the question your buyers ask AI.