Can I track agentic visitors on my website, and what counts as one?

Level 1: referrals

A person clicks a link inside ChatGPT, Perplexity, Claude or Copilot. The visit may arrive with a referrer such as chatgpt.com or perplexity.ai, and some links add utm_source. GA4 can analyze the sources; Plausible now groups known ones in its AI Assistants channel. Agent Tracking also shows identifiable assistant referrals separately. A missing referrer cannot be inferred reliably.

Level 2: fetches

Origin logs show requests claiming a crawler or user-triggered fetcher identity. Supported identities can be checked against current published address ranges; absent support remains unknown. Bursts group requests close in time but cannot reveal whether they answered a particular prompt.

curl -sS -X POST https://agenttracking.co/api/logs/example.com \
  -H "Authorization: Bearer wmt_your_token" \
  -H "Content-Type: text/plain" --data-binary @/var/log/nginx/access.log

Level 3: tool calls and in-browser execution

Supported browser tool APIs can be instrumented on the page; remote MCP tools require authenticated server reports. Technical call outcomes and browser goal attempts remain distinct from confirmed shop or booking receipts. Raw tool argument values are not retained.

<script defer data-domain="example.com" src="https://agenttracking.co/agent.js"></script>

Read the evidence together

Compare requested pages, technical tool failures and confirmed server receipts while retaining their separate denominators. Agent Tracking provides these views in an open-source, self-hostable product. Agent observability tools can also see visits if your own instrumented agent makes them.

In short