AI bot traffic analysis

See which AI bots are crawling your website.

AI crawlers visit your site daily, but Google Analytics can't see them. Finseo reads your logs and shows what GPTBot and ClaudeBot fetch.

Start free
halvard.com · Frankfurt

Latest request

LowHigh

500+ teams measure their AI visibility with Finseo

WürthKellogg'sPanasonicRaiffeisenJägermeisterDethleffsTrustpilot

See it

Every AI crawler visit to your site, in one log.

Finseo reads your logs and shows each request with the bot, the page, the status code, your response time and the city it came from.

Major crawlers are checked against their published IP ranges, so fake bots don't count.

Open any URL to see its visits, bots, problem rate and full user agents.

Latest visits · halvard.comLive
ChatGPT-User200182 ms
GET /jackets/storm-shell14:32:08 · Des Moines
Meta-ExternalAgent40486 ms
GET /guides/waxing14:31:55 · Luleå
Googlebot200214 ms
GET /sitemap.xml14:31:42 · The Dalles
Bingbot5031,240 ms
GET /cart?step=214:31:29 · Madrid
Latest visits · halvard.comLive
ChatGPT-User200182 ms
GET /jackets/storm-shell14:32:08 · Des Moines
Meta-ExternalAgent40486 ms
GET /guides/waxing14:31:55 · Luleå
Googlebot200214 ms
GET /sitemap.xml14:31:42 · The Dalles
Bingbot5031,240 ms
GET /cart?step=214:31:29 · Madrid

Understand it

Which bot reads which page, and why.

Each visit is sorted by what the bot does with your page. A block on the wrong bot keeps you out of answers.

01 Crawlers

The AI crawlers arrive

GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot and PerplexityBot, the major ones checked against their IP ranges.

Verified crawlers · halvard.com

GPTBotOpenAI IP verified
ChatGPT-UserOpenAI IP verified
ClaudeBotAnthropic IP verified
PerplexityBotPerplexity IP verified

Verified crawlers · halvard.com

GPTBotOpenAI IP verified
ChatGPT-UserOpenAI IP verified
ClaudeBotAnthropic IP verified
PerplexityBotPerplexity IP verified

02 Pages

They fetch your pages

Product pages, guides and filter URLs, with every visit and the status code per bot.

Visits per page

/jackets/storm-shell

142bot visits, last 30 days

ChatGPT-User64200
PerplexityBot41200
ClaudeBot37403

Visits per page

/jackets/storm-shell

142bot visits, last 30 days

ChatGPT-User64200
PerplexityBot41200
ClaudeBot37403

03 Purpose

Sorted by what they do

Live answer while someone asks, search index for AI search, or training for future models.

Visits by purpose · last 30 days

Live answer18 %
Search index47 %
Training35 %

Visits by purpose · last 30 days

Live answer18 %
Search index47 %
Training35 %

04 Blocked

A block, found and fixed

ClaudeBot gets a 403 from the firewall. Once the bot rule allows it, the page answers 200 again.

ClaudeBot

/jackets/storm-shell

Blocked by the firewall403

Readable again200

Allow ClaudeBot in the bot rule

ClaudeBot

/jackets/storm-shell

Blocked by the firewall403

Readable again200

Allow ClaudeBot in the bot rule

Where they come from

See where crawler requests come from.

The map colours each country by the bot visits from its data centres, or by your average response time. Slow regions show where crawlers wait on your server.

LowHigh

Countries of the bots' servers, mostly cloud data centres, not your visitors.

Countries · 30 days

The blind spot

Google Analytics shows no AI crawlers at all.

AI crawlers don't run JavaScript, so they never reach your analytics. Only your server logs show which pages ChatGPT, Claude and Perplexity read.

Cloudflare integration
halvard.com/guides/layering

Layering for a wet week in Scotland

Google Analytics0 bot sessions
PerplexityBotlast visit today, 14:32
Status200
PurposeSearch index
Data centreAshburn, US

Only in server logs

Why it matters

Bots read far more than they send back.

Cloudflare measured how many pages AI crawlers read for every visitor they send back. All the more reason to make sure they read the right pages.

Anthropic

38,000 : 1

pages crawled by Anthropic per visitor referred

OpenAI

1,091 : 1

crawls by OpenAI per visitor referred

Perplexity

195 : 1

crawls by Perplexity per visitor referred

Source: Cloudflare, The crawl-to-click gap, July 2025 data

Fix it

Fix what keeps crawlers away from your pages.

Blocked bots, dead URLs and wasted crawl budget, each with what it means and what to change.

Blocked crawlers

Blocked · 401 · 403 · 429

429

PerplexityBot

58

visits in 30 days

Your server answered “429 Too Many Requests” and turned the bot away. Perplexity can't read this page right now.

Check the rate limit for PerplexityBot

Crawl instructions

/robots.txt912
/sitemap.xml640
/entities.txt182

Content pages61.2 %

Crawl budget · all AI bots

Pages that no longer exist

Not found

404

/jackets/storm-shell-2023

38 bot visits in 30 days

Redirect it to the current page

Get started

Connect your logs in three steps.

Log sources

Cloudflare

Worker setup

Akamai

DataStream 2

Server-log API

REST endpoint

Log upload

nginx · Apache · .log

Fastly

Coming soon

AWS CloudFront

Coming soon

Step 01

Connect the source

A guided Cloudflare Worker setup, Akamai DataStream 2 or the server-log API keeps bot visits flowing in.

Step 02

Or upload logs

nginx, Apache or Cloudflare logs as .log or .txt, up to 1 GB per file. Finseo detects the format.

Step 03

Read the bots

Finseo recognises the AI crawlers, checks the big ones against their IP ranges and sorts each visit by purpose.

Bot traffic FAQ

Questions about AI bot traffic.

What customers say

Rated 5.0 — and ready for your security review.

Single sign-on, roles per project and data hosted in Germany, with a data processing agreement for every organisation.

Harald Fischl

“We’ve worked with Finseo for months and we’re delighted. You can tell it comes out of agency life: smart, clear and genuinely practical.”

Harald Fischl

Founder, Max Online

SSO via SAML

Login via your identity provider

Data in Germany

EU data residency, GDPR

Roles & audit trail

Every change logged

Bot traffic demo

Which AI bots read your website?

Enter your work email and we will show you in 30 minutes which crawlers visit your pages and what they come for.