trakily.
AI traffic

See what AI assistants send you, and what it's worth

ChatGPT, Claude, Perplexity, Gemini and other assistants get their own channel, with the visitors, customers and revenue each one brings. Add a small server snippet to also see which AI crawlers read your pages.

Channel: AI / LLMLast 30 days
AssistantVisitorsRevenue
ChatGPT9 customers310$880
Claude5 customers142$470
Perplexity2 customers96$190
Gemini1 customers58$79
ChatGPT58%1,840
Claude16%520
Perplexity13%410
Gemini (user fetch)8%260
AI referrals

AI assistants as a channel, with revenue

Works with the standard Trakily script. Nothing to set up.

ChatGPT
Claude
Perplexity
Gemini
Copilot
Meta AI
DeepSeek
Mistral
You.com

Recognised out of the box

Clicks from these assistants are classified as the AI / LLM channel, separate from search and social.

ChatGPT$880
Claude$470
Perplexity$190
Gemini$79

Revenue per assistant

Payments from your provider are matched to the visit that brought the buyer, so each assistant shows real revenue.

Asked ChatGPT for a toolMon 9:14
Read /pricingMon 9:16
Came back from GoogleWed 18:02
$Paid · Pro plan$169

The whole journey

Follow a customer from the first AI answer to checkout, even when they come back days later through another channel.

AI crawlers

Know which bots read your site

Optional: needs the server-side snippet, because crawlers don't run JavaScript.

ChatGPT58%1,840
Claude16%520
Perplexity13%410
Gemini (user fetch)8%260
Meta AI4%120
Mistral1%40

AI answers

An assistant fetched your page because someone asked about it right now: ChatGPT-User, Claude-User, Perplexity-User, Meta AI, Mistral, DuckDuckGo AI and Gemini.

Indexing

Search indexes that decide what gets cited: OAI-SearchBot, Claude-SearchBot and PerplexityBot, next to Googlebot, Bingbot and Applebot.

Training

Crawlers collecting data for models: GPTBot, ClaudeBot, Common Crawl, Bytespider, Meta, Google Vertex, Amazonbot, Cohere and more.

Grouped by page
/pricing820
/docs/getting-started610
/blog/stripe-attribution470
/390
/compare/plausible250

Group by page

See which pages AI crawlers fetch most, so you know what models and answer engines actually read.

AI assistant referrals + revenueStandard script
AI answer fetchesServer snippet
AI search & classic indexingServer snippet
Training crawlersServer snippet

What needs what

AI referrals come from the normal script. Crawler data comes only from the server snippet or a Vercel log drain.

Setup

A few lines on your server

Copy the snippet for your stack from Settings → Integrations → Bot tracking. It comes pre-filled with your website id.

Pre-filters on the user agent, so normal visitors never cause a request
Sends only the host, path, user agent and country. No cookies, no visitor IPs
Classified on our side against a list of known AI and search crawlers
proxy.ts
export function proxy(request: NextRequest, event: NextFetchEvent) {
  const userAgent = request.headers.get("user-agent") ?? "";
  if (BOTS.test(userAgent)) {
    event.waitUntil(fetch("https://trakily.com/api/bots", { … }));
  }
  return NextResponse.next();
}

Reports after the response is sent (event.waitUntil), so pages never wait on us.

FAQ

AI traffic, answered

No. Visitors who click a link in ChatGPT, Claude, Perplexity, Gemini and other assistants are recognised by the normal Trakily script and land in the AI / LLM channel, with revenue attributed like any other channel. The server snippet is only for crawlers.

Stop guessing. Start growing.

Add your site in two minutes and see which channels pay you by tonight.

14-day free trial No card required 2-minute setup