Skip to content

Bot IP verifier: is this really GPTBot, ClaudeBot or Googlebot?

This bot IP verifier checks an address against the IP lists AI and search crawlers publish, and the reverse DNS check Google and Bing document, so you know which hits are real.

Copy the address from your access log, firewall or CDN. Only the address is sent, never the log.

  • Free, no account
  • About 2 seconds
  • OpenAI, Anthropic, Google and 4 more
Example result

216.73.216.10

An address from Anthropic’s published list, checked against every bot

Run it on your site

216.73.216.10 belongs to Anthropic: ClaudeBot, Claude-SearchBot, Claude-User.

Verdict per bot

Whether 216.73.216.10 belongs to each bot, and the evidence
BotPurposeVerdictEvidence
ClaudeBotAnthropicModel trainingVerifiedInside 216.73.216.0/22 on Anthropic's published list (claude.com, dated 7 Oct 2026). The list covers all Anthropic crawlers, so it proves the operator, not this token. ClaudeBot list
Claude-SearchBotAnthropicAI searchVerifiedInside 216.73.216.0/22 on Anthropic's published list (claude.com, dated 7 Oct 2026). The list covers all Anthropic crawlers, so it proves the operator, not this token. Claude-SearchBot list
Claude-UserAnthropicFetches for a userVerifiedInside 216.73.216.0/22 on Anthropic's published list (claude.com, dated 7 Oct 2026). The list covers all Anthropic crawlers, so it proves the operator, not this token. Claude-User list
GPTBotOpenAIModel trainingFakeNot inside any range on OpenAI's published list (openai.com, dated 22 Sept 2026). GPTBot list
OAI-SearchBotOpenAIAI searchFakeNot inside any range on OpenAI's published list (openai.com, dated 2 Jan 2026). OAI-SearchBot list
ChatGPT-UserOpenAIFetches for a userFakeNot inside any range on OpenAI's published list (openai.com, dated 7 Oct 2026). ChatGPT-User list
PerplexityBotPerplexityAI searchFakeNot inside any range on Perplexity's published list (www.perplexity.com, dated 7 Feb 2025). PerplexityBot list
Perplexity-UserPerplexityFetches for a userFakeNot inside any range on Perplexity's published list (www.perplexity.com, dated 17 Oct 2025). Perplexity-User list
GooglebotGoogleSearch indexFakeGoogle documents a reverse-DNS check, and this address has no reverse DNS name. Googlebot list How to verify Googlebot
Google-ExtendedGoogleTraining and groundingFakeNot inside any range on Google's published list (developers.google.com, dated 7 Oct 2026). Google-Extended list
bingbotMicrosoftSearch indexFakeMicrosoft documents a reverse-DNS check, and this address has no reverse DNS name. bingbot list How to verify bingbot
Applebot-ExtendedAppleModel trainingFakeApple documents a reverse-DNS check, and this address has no reverse DNS name. Applebot-Extended list
CCBotCommon CrawlOpen training datasetFakeNot inside any range on Common Crawl's published list (index.commoncrawl.org, dated 11 Aug 2026). CCBot list

What the bot IP verifier checks

Each verdict comes from what the operator itself publishes, never from the user agent, which anyone can fake.

  • Published IP ranges

    OpenAI (GPTBot, OAI-SearchBot, ChatGPT-User), Anthropic, Perplexity, Google, Bing, Apple and Common Crawl publish their crawlers' address ranges. We fetch them, keep them 24 hours and show the range that matched.

  • Reverse DNS where it is documented

    Google, Bing and Apple also document a reverse DNS check. We look the name up, check it is in their domain and that it resolves back to the same address.

  • Verified, fake or unverifiable

    Verified when the address is on the list or passes the DNS check, fake when it fails the method the operator publishes, unverifiable when there is no method or the list is down.

  • One bot or all of them

    Pick the bot your log line claims to be, or check the address against every bot at once to see which operator it belongs to.

  • IPv4 and IPv6

    Both are checked, including IPv4 addresses written in IPv6 form (::ffff:66.249.66.1), as some servers log them.

  • The evidence, linked

    Every verdict names the range or the host name it rests on and links the operator's own list or documentation, so you can show it to your host or your firewall vendor.

Show 2 more checks
  • IPv4 and IPv6

    Both are checked, including IPv4 addresses written in IPv6 form (::ffff:66.249.66.1), as some servers log them.

  • The evidence, linked

    Every verdict names the range or the host name it rests on and links the operator's own list or documentation, so you can show it to your host or your firewall vendor.

How the bot IP verifier works

  1. Paste the address

    Copy an IP address from your access log, firewall or CDN, and pick the bot its user agent claims to be, or leave it on any bot.

  2. We ask the operators

    We check the address against each operator's published list and, where they document one, run the reverse and forward DNS lookup. Only the address is sent to us, never your log.

  3. Read the verdict

    Verified, fake or unverifiable for each bot, with the matching range or host name and a link to the source.

A list proves the operator, not the exact bot. Anthropic, for example, publishes one list for all its crawlers, so an address on it is an Anthropic crawler whichever token it sent.

Bot IP verifier questions

How do I check if an IP address is really GPTBot?

Paste it above with GPTBot selected. OpenAI publishes the address ranges GPTBot crawls from at openai.com/gptbot.json, and separate lists for OAI-SearchBot and ChatGPT-User. An address inside one of those ranges is OpenAI; anything else that sends a GPTBot user agent is fake. OpenAI documents no reverse DNS check, so the list is the only proof.

Does Anthropic publish ClaudeBot IP addresses?

Yes. Anthropic publishes one list at claude.com/crawling/bots.json and says an address on it is coming from Anthropic. The list covers all its crawlers, ClaudeBot, Claude-SearchBot and Claude-User, so it proves the request came from Anthropic but not which of the three it was. Anthropic documents no reverse DNS method.

Can I trust the user agent to tell which bot it is?

No. The user agent is a line of text the client writes itself, so any scraper can send "GPTBot" or "Googlebot". Only the address proves who sent a request: it must be inside the operator's published ranges, or, for Google, Bing and Apple, have a reverse DNS name in their domain that resolves back to it. The checker tests both, so a trusted name on an untrusted address shows as fake.

What should I do with a fake bot?

Block or rate limit it by IP address in your firewall, CDN or security plugin, not by user agent: blocking the user agent would also stop the real crawler. If many addresses fake the same bot, a rule that allows the bot only from its published ranges stops them all. Keep the real crawlers allowed if you want to appear in AI answers.

Why does a bot show as unverifiable?

Because its operator publishes no IP list and no reverse DNS method, so no one can prove an address is theirs, or because the operator's list could not be fetched just now. In the first case treat the user agent as a claim only; in the second, check again in a few minutes.

Is this bot IP verifier free?

Yes, with no sign-up. getReport is funded by donations, not plans. Only the address you paste is sent, never your log, and the full website report is free too.

All 59 tools →

Free, funded by the people who use it

No account, no paid plan. Donations keep it running.

Chip in