Fake Googlebot Checker - Verify Google, Bing & AI Crawler IPs

Paste IP and user-agent rows, or raw access-log lines, and find out which "Googlebot", "Bingbot" or "GPTBot" hits are real and which are spoofed. Each claim is checked against the vendor's official published IP ranges and, where the vendor documents it, forward-confirmed reverse DNS. Built for SEO log analysts, publishers that block AI crawlers, and security teams.

You also get a summary row with the share of bot-claiming hits that were spoofed.

Input examples

{ "rows": [ { "ip": "66.249.66.1", "userAgent": "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)" } ] }

{ "logLines": "66.249.66.1 - - [04/Oct/2026:10:00:01 +0000] \"GET /a HTTP/1.1\" 200 512 \"-\" \"Mozilla/5.0 (compatible; Googlebot/2.1)\"\n45.33.32.156 - - [04/Oct/2026:10:00:03 +0000] \"GET /c HTTP/1.1\" 200 512 \"-\" \"Googlebot/2.1\"" }

{ "logFileUrl": "https://example.com/logs/access.log", "onlyClaimedBots": true, "maxUniqueRows": 5000 }

Input guide

Field	What it does
rows	{ip, userAgent} objects, or raw log lines as strings.
logLines	Pasted Apache/Nginx combined-format lines, or JSON lines with ip and user_agent.
logFileUrl	URL of a plain-text log (up to about 60 MB).
aggregateByIp	One verdict row per unique IP + claimed bot, with a hit count (default on, recommended for logs).
onlyClaimedBots	Skip hits whose user-agent is a normal browser.
maxUniqueRows	Safety cap on unique pairs checked (default 2000).

Output

Field	Meaning
type	verdict, summary or error
ip, userAgent, claimedBot, hits	The input pair, which bot it claims to be, and how many log hits it represents
verdict	genuine, spoofed, unverifiable, not-a-known-bot, unverified (lookup failed) or invalid-ip
method	official-ip-list, fcrdns, ip-list+fcrdns, reverse-dns
ptr, reason	Reverse DNS name and a plain-language explanation
spoofedShare (summary row)	Percent of verifiable bot-claiming hits that were spoofed, plus byBot counts

Sample row:

{"type":"verdict","ip":"203.0.113.50","claimedBot":"Googlebot","hits":1,"verdict":"spoofed","method":"ip-list+fcrdns","reason":"not in the official IP list; no reverse DNS (PTR) record (ENOTFOUND)"}

Which bots are verified

• By official IP list: Googlebot (and Google special crawlers and user-triggered fetchers), Bingbot, GPTBot, ChatGPT-User, OAI-SearchBot, PerplexityBot, Perplexity-User, Applebot, DuckDuckBot, CCBot.
• By reverse DNS: Googlebot, Bingbot, Applebot, YandexBot (forward-confirmed); Baiduspider (reverse only, Baidu publishes no forward records).
• Reported as unverifiable: ClaudeBot and other Anthropic agents (Anthropic publishes no IP ranges), Meta crawlers, Bytespider, Amazonbot. The Actor says so instead of guessing.

Use it with

• Export a day of server logs, run this Actor, then block or rate-limit the IPs with verdict: spoofed.
• Pair with ai-search-readiness-audit: that one checks your robots rules, this one checks who is really crawling.

FAQ

Is a spoofed verdict certain? It means the IP is not in the vendor's current official list and (where applicable) fails reverse DNS. Lists are fetched live on every run, so a very new range can briefly be missed.

Why do some IPv6 hits pass? Official lists include IPv6 ranges and they are matched too.

Limits (honest)

• Verdicts depend on the vendors' published lists being reachable; if a list cannot be fetched the row says unverified, is not charged and is never marked genuine.
• Only the user-agent string decides which bot a hit claims to be.
• DNS checks use the platform resolver; a timeout shows as unverified.

Price

Pay per event: ip-verified $0.0005 per unique IP + bot pair with a verdict. Example: a log with 2,000 unique crawler IPs costs $1.00. Aggregation means a million hits from 2,000 IPs is still 2,000 events. Rows that could not be checked and the summary row are free.