Free Site Checkup

It runs ten of our free tools on your homepage and the site-wide files beside it, each exactly as that tool runs on its own page. Any address you enter is checked as its site: example.com/blog/post checks example.com.

• Redirects: the homepage over http and https, with and without www, and whether each address ends on a working page.
• robots.txt: whether the file blocks Googlebot from the whole site, and any line that breaks the file's rules.
• Sitemaps: whether robots.txt or a common address leads to a sitemap, and whether the sitemaps it lists answer.
• Title and meta tags: whether the homepage has a title and a meta description, and how long each is.
• Structured data: every JSON-LD, microdata and RDFa block, with its errors.
• Hreflang: the homepage's language versions and whether each one links back.
• AI crawler access: what robots.txt tells GPTBot, ClaudeBot, PerplexityBot and the other AI crawlers.
• Social share preview: what a shared link to the homepage shows, and whether its image loads.
• Favicon: whether an icon meets the rules Google uses for the icon in search results.
• llms.txt: whether the site publishes one, and whether it follows the format.

It uses one fixed order, the same for every site, and shows the first three problems it finds in that order. The order puts first what keeps the whole site out of search, then what stops pages being found and understood, then reach beyond classic search, and last the polish.

1. The site could not be reached
2. The homepage asks search engines not to index it
3. robots.txt blocks Google from the whole site
4. An address of the homepage ends in an error
5. No sitemap was found
6. A sitemap has errors
7. The homepage's structured data has errors
8. The homepage has no title or no meta description
9. The homepage's language versions have errors
10. robots.txt blocks AI search crawlers
11. Shared links to the homepage show no image
12. No icon meets Google's favicon rules
13. No llms.txt file

A noindex on the homepage outranks a missing sitemap because it removes the page from Google's results outright, while a missing sitemap only makes pages slower to find. A missing llms.txt comes last: the format is a proposal that some AI tools read, and no search engine requires it.

Because a single number hides what to do. Two sites with the same number can have nothing in common: one may block Google from every page, the other may lack a share image. The checkup counts the problems it found, names them, and links each one to the check that explains it.

The count adds up the problems each of the ten checks reports, the errors and warnings each would show on its own page. A finding one check repeats (the same missing image on every social platform, the same noindex on every address of the homepage) counts once in that check, but a problem two checks both see counts once in each: a noindex homepage is reported by the redirect, meta tag and hreflang checks alike. A check that could not finish is counted as nothing found, never as a pass and never as a problem.

It means the site did not give that check a full answer right now: it answered a server error, refused or rate-limited our checker, took longer than the check allows, or had already had its share of checks this minute. Such an answer says nothing lasting about the site, so the checkup marks the row "Could not check" and does not count it.

A row marked "Not run" was skipped: that check is temporarily unavailable, or the site could not be reached at all, in which case the other checks are not run. Run the checkup again in a few minutes, or open the check on its own page.

A row that says "Partly checked" got a definite answer with a real limit on the rest, and what it could read still counts. When the limit is our own (a robots.txt over Google's 500 KiB limit, a homepage too large to read whole), the report is complete. When it is about the site right now (a share image whose host refused us, a www address that did not answer), the report is marked partial.

A report with any row that could not finish, was not run, or was partly checked for a reason about the site right now is marked partial. It is not kept for comparison, so the next run's "what changed" never reports an outage as a change on your site.

Work down the list from the top. Each row has a link that opens that tool with your site filled in, where you can see the full result: the deciding robots.txt line, the redirect chain hop by hop, the schema error and the field it is on.

When you have fixed something, run the checkup again. Your recent checkups stay in this browser, and the next one on the same site shows what changed since the last, such as a problem that is gone or a new one. You can also share a report link with whoever makes the fix; the link works for 30 days.

The checkup reads one page. Problems such as a stray noindex, a broken canonical or a missing title can sit on any page of a site, so for a full picture you need a crawl of every page.