Observe requests, not promises.
A recorded crawler request can show that a supported identity reached the tracking layer and requested a URL. It does not prove indexing, training use, a citation, or a human visit.
See which AI crawlers reach your website.
Monitor supported GPTBot, ClaudeBot, PerplexityBot, and search-crawler requests, the public URLs they request, and the responses they receive.
Ready to turn AI visibility into a workflow?
Open The AI WorkspaceConnect activity to the next check.
Use URL coverage and response outcomes to investigate access, discovery, and internal-link gaps, then verify changes in the full TruboRankAI workflow.
AI crawler analytics guide
AI Bot Tracker evidence for GPTBot, ClaudeBot, PerplexityBot, and search crawlers
Use crawler request evidence to diagnose public-page coverage and failures while keeping crawl activity separate from mentions, citations, referrals, and conversions.
What the AI Bot Tracker records
The tracker organizes supported server-side request evidence such as time, requested public URL, response status, declared crawler identity, and the identity confidence available to the integration. Server-side collection matters because crawlers do not need to execute browser analytics.
Coverage depends on where the tracker is installed and where the response is served. A CDN cache, firewall, proxy, or separate static host may handle a request before it reaches application code. Treat an empty report as no matching event observed in the selected collection path, not proof that no crawler visited any layer.
How to interpret crawler activity
Start with response outcomes. A 200 response can show successful delivery at the observed layer; a 403 points to an access decision; a 404 points to a missing route; redirects need the final destination. Group requests by canonical priority page so repeated parameters and assets do not hide the pages that matter.
Keep provider purposes separate. OpenAI documents distinct controls for GPTBot and OAI-SearchBot, and other providers publish their own crawler identities. Preserve the exact observed identity instead of combining every AI-related user agent into one total.
A crawler hit is not a ranking, mention, citation, recommendation, or human referral. Review those outcomes in Search Console, answer observations, analytics, and conversion systems with their own dates and definitions.
From tracking evidence to action
Investigate failed requests first, then important pages with no observed coverage. Check robots rules, authentication, firewall decisions, response status, sitemap discovery, and internal links without weakening private-route protection. A manual test can verify delivery but cannot force a provider to crawl.
Use TruboRankAI as a connected workflow: start with public discoverability checks, add AI Bot Tracking for supported owned-site evidence, improve AEO and GEO where the page itself is weak, and rescan after deployment. The platform keeps these evidence layers separate so one positive signal is not reported as a visibility guarantee.
Free checks versus the full platform
| Need | Best starting point | Next step |
|---|---|---|
| Check whether a crawler is allowed | AI Bot Access Checker | Review the exact public URL and policy |
| Observe supported crawler requests | AI Bot Tracker | Compare URL coverage and response outcomes |
| Diagnose the whole website | AI Discoverability Platform | Prioritize, implement, rescan, and measure |
Frequently asked questions
What does an AI bot tracker prove?
It can prove that a matching request was observed at the configured collection layer. The recorded status and identity evidence determine what can be concluded about delivery and crawler classification.
Can AI Bot Tracking guarantee ChatGPT or Perplexity mentions?
No. Crawler requests, answer mentions, citations, human referrals, and conversions are separate events and should be measured separately.
Why might the tracker show no activity?
There may have been no matching request, or the request may have been served by a CDN, firewall, proxy, static host, or route outside the tracker. Validate collection coverage before changing access rules.
Related internal links
Methodology and limits
Reviewed on 2026-10-22 against the implemented TruboRankAI server-side tracking workflow and current provider crawler documentation. Observed requests, declared identities, verification confidence, citations, referrals, and conversions remain separate evidence types.