Check that an IP address really belongs to Googlebot, Bingbot or GPTBot

How to use the tool

Paste one IP address per line, or lines straight from your server logs: the tool pulls out the IPv4 and IPv6 addresses and removes duplicates. Then click “Check”. Each address gets a verdict, the operator’s name when it belongs to one, the range of the official list that contains it and the name returned by its reverse DNS.

Below the table, the lists checked give, for each bot, the number of address ranges and the publication date stated by the operator. The tool downloads them from their official addresses and keeps them for 24 hours; a list that does not answer keeps its last known version.

Why check a crawler’s address

The name a crawler gives in its requests, the user agent, anyone can copy. A log line that says “Googlebot” therefore proves nothing: Google itself explains how to check that a crawler using its name really is its own.¹ Bing publishes the same advice for Bingbot.²

The check matters as soon as you read your logs: counting Googlebot’s real visits, spotting a bot that poses as it, or opening a firewall to the crawlers you want to let in without opening it to their imitators.

The two ways to verify

The first is to look the address up in the lists the operators publish, in JSON for most of them. Google publishes five, one per family of crawlers, and states that the address can be matched against these ranges.¹ Microsoft has published Bingbot’s since November 2021,³ OpenAI one per bot,⁴ and so does Perplexity.⁵

The second goes through reverse DNS, confirmed by forward DNS. You ask for the name tied to the address, check that it belongs to the operator’s domain, then ask for the address of that name: it must match the starting address.¹ The round trip matters, because whoever owns an address can write anything in its reverse DNS, but not in the DNS of googlebot.com. Google, Bing, Apple and Common Crawl document this method.²⁷⁸

The tool runs both methods on every address. An address is verified if it appears in an official list, or if its reverse and forward DNS agree on a domain documented by the operator.

What each operator publishes

Not every operator offers the same ways to verify its crawlers. Here is what the tool checks, based on each one’s documentation read on October 18, 2026.

Anthropic states that a request from an address on its list does come from Anthropic, but advises against blocking its bots by IP address: to refuse crawling, it points to robots.txt.⁶ Other crawlers, such as those of Meta or Mistral AI, have no list tracked by the tool.

Reading the verdict

Verified: the address belongs to the operator shown, through its official list, through DNS, or both. The details name which of the operator’s bots is concerned when its list says so; Anthropic’s list, shared by its three bots, cannot tell them apart.

Spoofed name: the reverse DNS announces an operator’s domain, for example a googlebot.com name, but that name does not resolve back to the address. Any address owner can write such a name; only the operator’s DNS counts.¹

Not a listed bot: the address is in none of the lists and its DNS leads to no tracked operator. If your logs present it as Googlebot, GPTBot or another bot on this page, the bot name is spoofed. A private or local address is flagged separately, since it cannot come from the internet.

What the tool cannot know

Lists change: operators update them at their own pace, and the tool rereads them at most every 24 hours. The publication date of each list appears in the result. Bingbot’s is dated January 3, 2024 at the time of writing: a recent Bing address may be missing from it, and reverse DNS then confirms it. Amazon’s pages give no date.

The tool checks ten addresses at most at a time, and accepts 20 checks per 10-minute window and per visitor, across all the site’s tools. The addresses entered are not stored. It does not read whole log files: to count a bot’s visits over thousands of lines, paste a sample of the addresses that come up most.

After the check

Knowing who visits the site leads to deciding who may crawl what. The AI crawler access test tells you, bot by bot, what your robots.txt allows, and the robots.txt generator writes the rules to set. When the logs show a Googlebot that rarely comes, or always to the same pages, that is a crawling issue I handle in the SEO support I provide to my clients.

Frequently asked questions