A user agent is a claim, not evidence.
Anyone can send a request that says GPTBot. RankSage checks each one against the DNS records the crawler's operator publishes, and reports verified, rejected and unverifiable separately.
The problem
Every other AI crawler report counts user-agent strings. Scanners routinely wear crawler names, so those reports quietly tell you that ChatGPT has been reading pages it never touched.
Verified AI Crawlers on one screen.
What you see in RankSage.
Verification verdicts, as they appear in the product.
Address resolves to the operator's published records
Claimed GPTBot, resolved to an unrelated host
Operator publishes no records to check against
Unverifiable is its own bucket on purpose. It is never counted as a confirmed visit, and never dropped either.
What it measures, and what comes out.
Outcome and evidence at each step, never the recipe. Every number here can be traced to something stored.
- 01
The claim is checked, not accepted
Each hit's address is resolved back against the records the operator publishes for its crawlers. Operators publish these precisely so they can be verified.
- 02
Three outcomes, kept apart
Verified, rejected, and unverifiable. The third exists because some operators publish nothing to check against — treating that as a pass would be a lie, and dropping it would be a different one.
- 03
Rejected traffic is shown, not deleted
Requests that failed the check are kept and reported in their own bucket. Knowing that a hundred requests wore ClaudeBot's name and failed is useful information about your site.
- 04
Then the useful question
Once you can trust the numbers: which engines read which pages, how often, and whether your own rules are blocking any of them without your knowledge.
- 05
Crawls, set against the visits they earn
For each AI operator, the pages it crawled sit beside the referral visits it actually sent you — so you can see which engines read a lot and send little, and which quietly convert reading into traffic.
What it does not do.
Disqualifying fast is part of the product. These are the edges, written down.
It needs a small snippet on your server or edge, or the WordPress plugin. A browser script cannot see crawlers, because crawlers do not run scripts.
Where an operator publishes no verification method, RankSage records the visit as unverifiable. It will not guess.
The address is used for the check and discarded. RankSage stores the verdict, never the address.
The capabilities it joins.
Read the longer version.
Questions about Verified AI Crawlers, answered.
What it measures, what it needs, and where its edges are.
Check the claim. RankSage resolves each hit's address against the DNS records the crawler's operator publishes and records the result as verified, rejected or unverifiable — so spoofed requests never inflate your totals.
Trivially. A user agent is a self-declared string, and credential scanners routinely wear crawler names. This is exactly why verification exists.
No. The address is used to run the check and then discarded. What is stored is the verdict.
AI is already answering for you.
Sign in to Google Analytics and Search Console. Read-only, about five minutes, no code on your site for that step.
No payment details. One email when access opens.