Shipped: Agent Identity & Human Trust Layer
WebDecoy now verifies AI agents with Web Bot Auth (RFC 9421), catches agent impersonation, and grades human trust on clearance tokens. Everything that shipped.
bot-detectionWebDecoy now tracks visitors arriving from ChatGPT, Perplexity, Gemini, and 8 other AI platforms. Measure the traffic channel your analytics is missing.
WebDecoy Team
WebDecoy Security Team
Google Analytics shows you search traffic. It shows you social traffic. It shows you direct traffic. But when someone asks ChatGPT about your product, clicks the link in the response, and lands on your pricing page — what does your analytics show?
“Direct.” Or worse, nothing at all.
That changes today. WebDecoy now tracks LLM referral traffic across 11 AI platforms, giving you visibility into the fastest-growing acquisition channel on the web.
AI search is eating traditional search. Gartner projects that traditional search volume will drop 25% by 2026 as users shift to AI-powered alternatives. When your potential customers ask Perplexity “best bot detection tools” or tell Claude “compare WebDecoy vs Cloudflare bot management,” the AI cites your content and links to your site.
But here’s the problem: traditional analytics tools weren’t built for this. Most AI platforms strip or obscure referrer data. Google Analytics lumps AI-referred visits into “direct” or “(not set)” buckets. You’re flying blind on a traffic channel that’s growing exponentially.
Without LLM referral tracking, you can’t answer basic questions:
WebDecoy’s detection script checks the document.referrer header on every page load and matches it against a maintained list of known AI platform domains. When a match is found, a lightweight beacon fires with the llm_referral flag.
Key technical details:
llm_referral — Categorized separately from bot detections, scraper alerts, and other signals.This is an analytics feature, not a security one. You’re not blocking anything — you’re measuring a traffic source that was previously invisible.
WebDecoy identifies referral traffic from every major AI platform:
| Platform | What It Covers |
|---|---|
| ChatGPT | ChatGPT web app, shared conversations, search links |
| Perplexity | Perplexity search results and answer citations |
| Google Gemini | Gemini app responses and citations |
| Claude | Claude.ai conversation links |
| DeepSeek | DeepSeek chat and search responses |
| Microsoft Copilot | Copilot web responses and Bing AI integration |
| You.com | You.com AI search results |
| Phind | Phind developer search responses |
| Kagi | Kagi AI search assistant |
| Meta AI | Meta AI responses across Meta platforms |
| Grok | Grok responses on X (Twitter) |
As new AI search platforms emerge, the referrer list is updated — no action required on your end.
The AI Referrals dashboard gives you four views:
Top Referring AI Platforms — Which AI services are sending you the most traffic. If Perplexity drives 3x more referrals than ChatGPT, your content strategy should account for that.
Traffic Trends Over Time — Are AI referrals growing week over week? This is the growth curve you need to track, especially as AI search adoption accelerates.
Top Landing Pages — Which pages receive the most AI-referred traffic. If your pricing page gets heavy Perplexity referrals but your docs don’t, that tells you what AI models are citing — and what they’re not.
Volume Breakdown by Platform — Proportional view across all tracked platforms. Useful for understanding which AI ecosystems your content performs best in.
When someone asks ChatGPT “how do I detect AI scrapers on my website” and ChatGPT links to your blog post, that’s a lead. It’s no different from a Google search click — except it’s been invisible until now.
Early data from publishers tracking AI referrals shows that AI-referred visitors often have higher engagement metrics than organic search visitors. The hypothesis is straightforward: a user who asked a specific question and got pointed to your content has strong intent.
If you’re investing in content marketing, you need to know whether AI platforms are citing your content. Without referral data, you’re optimizing for Google rankings alone while ignoring a channel that Similarweb data shows is growing double digits month over month across major AI search platforms.
LLM referral data tells you:
WebDecoy already detects AI scrapers (GPTBot, ClaudeBot, and 20+ crawlers) and identifies RAG bots fetching content in real-time. LLM referral tracking completes the picture by capturing what happens after the AI processes your content — the actual human visit.
The three signals together tell the full story:
AI SCRAPER → AI model ingests your content
RAG BOT → AI fetches your content for a user query
LLM REFERRAL → Human clicks through from AI responseYou can now track the entire AI content lifecycle: from crawl, to citation, to click.
LLM referral tracking is available now for all WebDecoy customers. If you already have the WebDecoy detection script installed, referral data is being collected automatically — no configuration changes needed.
To view your data:
For API access, query referrals with:
GET /v1/detections?flags_contain=llm_referralFull documentation is available in the AI Detections docs.
LLM referral tracking is the foundation for deeper AI traffic analytics. We’re building toward:
AI search is rewriting how people discover products and content. The companies that measure this channel early will have a significant advantage over those still treating AI traffic as “direct.”
Already using WebDecoy? Your LLM referral data is waiting in the dashboard. Not yet a customer? Start tracking your AI traffic today.
WebDecoy now verifies AI agents with Web Bot Auth (RFC 9421), catches agent impersonation, and grades human trust on clearance tokens. Everything that shipped.
bot-detectionWebDecoy is now live on the WordPress.org plugin directory. One-click install, automatic updates, and 100% local bot protection with no API key required.
bot-detectionGoogle signs crawler traffic with Web Bot Auth (RFC 9421). What signed bots mean for detection, verified-bot allowlists, and how WebDecoy verifies them.
bot-detectionLike this post? Share it with your friends!
Get a personalized demo from our team.