# Is that really Googlebot? How to spot an impostor

> Anyone can call their scraper "Googlebot" or "GPTBot". Here's how to check a crawler is who it says it is, with the reverse DNS and IP-range methods Google, OpenAI and others document.

Punchline Analytics · 2026-11-16

**Short answer:** never trust the user agent alone. A real Googlebot comes from an IP address whose reverse DNS name ends in `googlebot.com`, `google.com` or `googleusercontent.com`, and that name resolves back to the same address; or from an address in the IP ranges Google publishes. Other crawlers (GPTBot, ClaudeBot, PerplexityBot and more) can be checked against their operators' published IP ranges. A request that claims a famous name from anywhere else is an impostor.

## Why it matters

Scrapers borrow well-known crawler names because sites tend to let them in. If you allow "Googlebot" past a paywall, a rate limit or a bot filter on the strength of its user agent, you've let in anyone who types that string. And if your analytics counts impostors as Google, your crawl numbers are wrong.

## Method 1: reverse DNS, then forward DNS

This is [Google's documented method](https://developers.google.com/crawling/docs/crawlers-fetchers/verify-google-requests) for checking a single request:

1. **Reverse lookup** the requesting IP address: `host 66.249.66.1`.
2. **Check the name** ends in `googlebot.com`, `google.com` or `googleusercontent.com`.
3. **Forward lookup** that name: `host crawl-66-249-66-1.googlebot.com`.
4. **Check it resolves to the same IP** you started with.

Step 3 is the one people skip, and the one that matters: anyone who controls an IP address can set its reverse DNS name to anything, including something ending in `googlebot.com`. Google [made the same point back in 2006](https://developers.google.com/search/blog/2006/09/how-to-verify-googlebot). Only the forward lookup proves Google controls the name.

## Method 2: published IP ranges

For checking lots of requests, look the address up in the lists operators publish instead of doing DNS lookups each time:

- **Google** publishes its crawlers' IP ranges as JSON, now at [developers.google.com/crawling/ipranges/](https://developers.google.com/search/blog/2026/03/crawler-ip-ranges) (the old location will redirect).
- **OpenAI** publishes ranges per crawler, such as `openai.com/gptbot.json` and `openai.com/searchbot.json` ([docs](https://developers.openai.com/api/docs/bots)).
- **Perplexity** publishes JSON ranges for each of its crawlers ([docs](https://docs.perplexity.ai/docs/resources/perplexity-crawlers)).

Refresh the lists regularly; they change. The [AI agent directory](https://punchlineanalytics.com/agents) links each crawler's list ([also as JSON](https://punchlineanalytics.com/agents.json)).

## Method 3: signatures

Some agents now sign their requests (Web Bot Auth: an HTTP message signature checked against keys the operator publishes). A valid signature is the strongest evidence of all, because it can't be borrowed by copying a string. Where an agent signs, check that first.

## What to do with impostors

- **Don't give them a crawler's privileges:** no paywall bypass, no relaxed rate limits.
- **Count them separately.** They're scrapers, often aggressive ones, and lumping them in with the real crawler hides both.
- **Don't panic about them in analytics,** as long as your analytics labels them. Punchline checks every request that claims a known crawler's name against its operator's published addresses (and signatures where there are any) and labels the ones that fail as **impostors**, with the evidence. [See who's borrowing names on your site](https://app.punchlineanalytics.com/signup).
