Citare Tools · Free
AI Bot UA Detector
Paste any User-Agent string (or a whole access-log line) — we match it against 25+ AI crawler patterns and tell you who’s crawling, what they do with the data, and whether their IPs can be verified.
Free. No signup. Cached 24h.
Frequently asked
What is the GPTBot user-agent string?
GPTBot's User-Agent string is "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.x; +https://openai.com/gptbot" — versions step periodically. It's the user-agent OpenAI uses to crawl public web content for training future GPT models. Other related OpenAI user-agents include ChatGPT-User (the agent making fetches when a ChatGPT user clicks a link or a custom GPT browses the web) and OAI-SearchBot (the indexer used for ChatGPT's live web search). All three are different bots with different roles — paste any UA into Citare's detector to see which one you're looking at.
How do I identify which AI bot is in my server access logs?
Copy the User-Agent string from the log line and paste it into the form above. Citare's AI Bot UA Detector accepts either a bare UA string ("GPTBot/1.2") or a full Apache/nginx combined-log line — it extracts the UA portion automatically. The detector then runs the string against 25+ AI crawler patterns and returns the bot name, the provider (OpenAI, Anthropic, Google, Perplexity, etc.), the bot's role (training / live search / agentic), and whether the bot's IPs can be verified against a published IP-range list.
What is the difference between GPTBot, ChatGPT-User, and OAI-SearchBot?
They're three different OpenAI user-agents with three different jobs. GPTBot crawls public content for training future GPT models (block this if you don't want OpenAI to train on your content). ChatGPT-User fetches a URL on behalf of a ChatGPT user when they click a link or a custom GPT browses a site (block this if you don't want individual ChatGPT users to see your content via in-app browsing). OAI-SearchBot indexes web content for ChatGPT's live web search feature (block this and you disappear from ChatGPT's grounded citations). Most sites should allow OAI-SearchBot at minimum — that's the bot that controls ChatGPT citation visibility.
Does this UA detector send my data anywhere?
No. Citare's AI Bot UA Detector runs entirely in your browser. The bot definitions and pattern-matching logic are bundled into the page's JavaScript, and detection happens client-side without any network calls — nothing leaves your machine. Pasting raw access-log lines is safe; we don't see them, we don't log them, and we don't cache them. There's also no signup, no account, and no rate limit, because there's nothing on our backend to rate-limit.
Does Google have a separate AI bot user-agent for Gemini?
No published one. Google has Google-Extended (which controls Gemini and Vertex AI training opt-in) and GoogleOther (a generic Google-AI fetcher), but Gemini's live-grounding fetcher uses a stealth Chrome-style User-Agent rather than a published bot UA. That means Gemini's live grounding access is governed by Googlebot rules and JavaScript-rendered HTML access, not a dedicated GeminiBot string. If you want Gemini to ground answers from your site, treat it as a Googlebot question (allow Googlebot, ensure JavaScript-rendered HTML is parseable) rather than an AI-bot allowlist question.