Can AI assistants find your site? See what they see.
A free check of what ChatGPT, Google’s AI Overviews, Copilot, Claude and Perplexity can reach and read on any page: crawler access, content that loads without JavaScript, and structured data. Scored out of 100, weighted by what the engines themselves publish.
Checks the page you enter, plus the site’s robots.txt, sitemap and llms.txt. Nothing is stored.
What the engines themselves say
Most advice on AI visibility is guesswork. These are the few things the engines and crawler studies have put on record, and the checker is built on them.
On the record
Published by the enginesQuoted from the engines’ own documentation and one large crawler study, not measured here.
0
JavaScript runs found in more than 500 million GPTBot fetches. ClaudeBot and PerplexityBot behaved the same way: what isn’t in the HTML isn’t read.
Source · Vercel and MERJ crawler analysis
1
OpenAI crawler decides whether a site appears in ChatGPT search: OAI-SearchBot. Blocking GPTBot only opts out of training.
Source · OpenAI, Overview of OpenAI crawlers
0
special files or schema needed to appear in Google’s AI Overviews and AI Mode. The page has to be indexed and allowed a snippet.
Source · Google Search Central, AI features and your website
So the score follows the same order
Can a crawler reach the page, can it read the page without running scripts, and can it tell what the page is. Anything else is a bonus.
Read Google’s guidanceWhat no checker can see
- Whether an assistant already cites you: that depends on what your pages say, and who else mentions you
- Pages other than the one you enter: check a category and a product page too
- Whether the content answers the questions buyers actually ask
- What happens after a click: speed, checkout and trust, which the audits cover
How the score is worked out
Weighted by what decides visibility
Sixteen checks in four groups. A check that passes scores in full, a partial pass scores half, and the total is out of 100.
- 40 points
- Reach
- 35 points
- Read
- 20 points
- Understand
Each check, and why it carries the weight it does:
- Crawler access, 25. robots.txt is read the way each crawler reads it: OAI-SearchBot for ChatGPT search, Googlebot for AI Overviews, Bingbot for Copilot, Claude-SearchBot and PerplexityBot.
- noindex and nosnippet, 10. Either one keeps a page out of AI answers, however good it is.
- HTTPS, 5. The page has to answer, securely, at the address people use.
- Content without JavaScript, 15. The largest single check, because AI crawlers read the raw HTML and nothing a script adds later.
- Heading, title, description and canonical, 20. The plain signals every engine uses to say what a page is and which address is the real one.
- Structured data and identity, 15. JSON-LD that parses, and something machine-readable that says whose site it is.
- Sitemap, 5. So a crawler can find the rest of the site.
- llms.txt, 5. Optional. Worth having, but no major engine says it reads it.
It checks the one page you enter, so try your home page, a category and a product page. It can’t tell you whether an assistant already cites you: that depends on what the page says and who links to it.
The crawler rules come from the engines’ own documentation. OpenAI’s crawler list
Why llms.txt counts for five points
Google’s own guidance: “You don’t need to create new machine readable files, AI text files, or markup
to appear in these features.” llms.txt is a sensible, cheap file, and some tools read it. But no major search engine has said it uses the file to decide what to show, so a site with a perfect llms.txt and its content hidden behind JavaScript is still invisible.
Fix what the crawlers actually read first. Add llms.txt after.
When a score isn’t enough
Find out why, and what it’s costing
The checker reads one page from the outside. A health check on AI visibility reads the path a customer takes, and the audits above it read every template, then the code that renders them.
Fixed prices, excluding VAT. The complete audit is quoted.