The AI visibility panel is on by default for a new share link. It is the part people share, so it would be an odd thing to hide behind a setting nobody finds.
Turn it off in the link's settings if you would rather not include it.
What a viewer sees when it is on
The same panel you see, over the range they select:
AI answers / Indexing / Training / AI referrals — four cards, picking what is counted
Crawlers — views per operator (OpenAI, Anthropic, Google, Amazon, …)
AI pages — which of your paths the crawlers fetched
Plus the verification line: the share of crawler visits confirmed as genuinely from that operator.
Before you share it publicly
AI pages names your paths. It is the same disclosure as the Top pages card — your URL structure, listed. If you turned Top pages off for privacy reasons, turn this off too, or you have disclosed the same thing through a different card.
The honesty problem when you share it
This is the section worth reading before you screenshot a number into a thread.
On a browser-script-only install the AI panel undercounts, and a viewer has no way of knowing that. Most AI crawlers request raw HTML without running JavaScript, so the script never loads and the visit is never recorded. What the panel shows is then a floor.
You know that because you read How reliable is it. Someone opening your share link has not.
So if you are publishing these numbers — in a post, a pitch, a report — say which of the two you are publishing.
If only the script is reporting:
"Crawler views recorded by TrueStat's browser script. Crawlers that fetch raw HTML without running JavaScript are not included, so this is a lower bound."
If your server reports them too, say that instead — the undercount is much smaller and calling the figure a floor understates your own instrument:
"Crawler views recorded by TrueStat, from the page script and from our server's own request log."
Either way it is a stronger framing than an unqualified number: it says you know what your instrument measures.
The comparison that always holds is relative — which operators, which pages, which direction over time. Those survive a constant undercount. The absolute count does not.
What a viewer cannot tell from the panel
They cannot tell what you excluded. A crawler family you have no traffic from and one you filtered out look the same — empty.
They cannot tell how much is missing. Neither can you. Nobody can put a number on the undercount, which is exactly why the framing above matters.
The day you add server-side reporting
Server-side reporting captures the crawlers a browser script cannot see. The day you switch it on, your crawler numbers step up.
If you published a figure before then, expect the newer one not to match. That is missing traffic becoming visible, not growth — and it is worth saying so alongside the new number, for the same reason it is worth qualifying the old one.