GeoPromptTracker

Free tool — no sign-up

AI Bot Traffic Log Analyzer

Everything runs in your browser — log lines are never uploaded to our servers.

AI bot traffic report

Nothing yet — run the tool above.

Your robots.txt says what AI bots may do — your access logs show what they actually do. This analyzer parses pasted log lines entirely in your browser and reports which AI crawlers are visiting, how often, and which pages they're fetching.

Why it matters for AI search: AI crawler traffic is the ground truth of your GEO efforts. If OAI-SearchBot and PerplexityBot are actively fetching your pages, your content is being indexed for AI answers; if only training bots like GPTBot and CCBot show up, you're feeding models without gaining search-style visibility; if nothing shows up at all, your content isn't on AI platforms' radar yet — three very different situations demanding different responses, and indistinguishable without looking at logs. The per-path breakdown adds another layer: seeing which pages AI bots prioritize tells you what they consider your most valuable content.

How this tool works: paste raw access-log lines — Apache or nginx combined format works best, but any format containing user-agent strings is fine — and the analyzer matches each line against the same continuously updated AI bot registry that powers our other crawler tools: GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-User, Claude-SearchBot, PerplexityBot, Perplexity-User, Google-Extended, CCBot, Bytespider, Applebot-Extended, Amazonbot, Meta-ExternalAgent, and cohere-ai. You get total hit counts per bot, each bot's operator and purpose (training vs. search), and the top paths each bot requested. Everything runs client-side: your logs never leave your browser, which matters since access logs contain visitor IP addresses.

Limitations: the analyzer matches known user-agent substrings, so bots spoofing browser user agents or crawlers not yet in our registry won't be counted — the registry is refreshed monthly, but brand-new crawlers can lag a few weeks. Browser-based parsing also has practical size limits: paste up to a few hundred thousand lines comfortably, but multi-gigabyte log files should be pre-filtered (e.g. grep -i "gptbot\|claudebot\|perplexity" access.log) before pasting. And remember that a log sample is a snapshot — AI crawler activity is bursty, so trends across weeks are more meaningful than any single day's count. To change what these bots are allowed to do, pair this tool with the robots.txt Generator for AI Bots.

Frequently asked questions

Are my log files uploaded to your servers?

No. This tool runs entirely in your browser — the log lines you paste are parsed locally with JavaScript and never leave your machine. You can verify this in your browser's network tab.

What log formats are supported?

Any format where the user-agent string appears in the line works for bot counting — Apache and nginx combined log format additionally get per-path breakdowns, since the request path can be extracted from the quoted request segment.

How do I get my access logs?

On most hosts: cPanel's Raw Access Logs, nginx's /var/log/nginx/access.log, Apache's /var/log/apache2/access.log, or your CDN/hosting dashboard's log export (Cloudflare, Vercel, and Netlify all offer log access on paid tiers).

What does it mean if I see lots of GPTBot hits?

GPTBot is OpenAI's training crawler — heavy traffic means your content is being collected for model training. Whether that's good or bad depends on your stance; see our guide on the trade-offs. Search bots like OAI-SearchBot hitting you is generally the visibility signal you want.

Why do I see zero AI bot hits?

Either AI crawlers genuinely haven't visited (common for new or low-authority sites), your robots.txt blocks them, or your log sample is too small — AI crawler visits are bursty, so analyze at least a few days of logs before concluding anything.

Related