See what's read
Crawler hits per page and per bot, including error responses, so you can spot pages AI crawlers can't reach.
AI assistants now send visitors and read your pages. GoTrace counts both: the people who click through from ChatGPT, Claude, Perplexity, Gemini and Copilot, and the AI crawlers that read your site before an assistant quotes it.
Most analytics tools file ChatGPT under generic referrals, or lose it in "Direct". Here, every visit from an AI assistant lands in a dedicated AI channel, recognised from the referrer and from the utm_source tags assistants add to links.
Matched by referrer domain and by the utm_source tag on the link.
GPTBot, ClaudeBot and PerplexityBot don't run JavaScript, so no script-based analytics, ours included, can count them in the browser. Your web server sees every one of them. Send us its access log and we'll show you which AI crawlers read which pages, and how often.
# Every hour: send AI crawler hits from nginx 0 * * * * grep -iE 'GPTBot|ClaudeBot|PerplexityBot|…' \ /var/log/nginx/access.log | tail -n 20000 | \ curl -s -X POST \ -H 'Authorization: Bearer YOUR_API_KEY' \ -H 'Content-Type: text/plain' \ --data-binary @- \ https://gotrace.site/api/v1/sites/SITE_ID/crawler-logs
Your dashboard generates this line with the full crawler list, your site ID and the right URL.
When someone asks an assistant for a recommendation, the answer is built from pages its crawler has read. Knowing which of your pages get read, and which ones send buyers back, tells you where to put your effort.
Crawler hits per page and per bot, including error responses, so you can spot pages AI crawlers can't reach.
AI referral landing pages show which content assistants actually link to when they answer.
Compare AI visitors' conversion with search, then improve the pages that already earn AI traffic.
Some assistant apps open links without passing a referrer. ChatGPT usually adds utm_source=chatgpt.com to the links it shows, and we use that tag as well as the referrer, so these visits land in the AI channel instead of Direct.
No. Crawlers and bots are filtered out of your visitor statistics and counted separately, so they never inflate your traffic or skew conversion rates.
The standard "combined" access log format used by nginx and Apache by default, plain or gzipped. Lines from the last 30 days are counted.
That's your call. Blocking them in robots.txt keeps your content out of AI answers, but it also stops assistants from recommending you. Measuring first tells you what you'd be giving up.
Install the script, upload one log, and see your AI traffic today.
Start your free trial