How to view a page as Googlebot

Four ways, and what each one proves

1. Change the user agent in your browser

In Chrome: DevTools → three dots → More tools → Network conditions → turn off “Use browser default” and pick Googlebot. Reload the page.

The site now sees Googlebot’s name, and everything else about you unchanged: your address, your country, your cookies. That makes this the cheapest method and the weakest one. It catches a page that switches on the user agent alone. A page that checks the visitor’s IP address ignores your disguise completely and serves you the ordinary version, which is exactly how most serious cloaking is built.

Useful as a first look. Never trust a clean result from it.

2. URL Inspection in Search Console

For a site you own and have verified, this is the honest answer: Search Console fetches the page from Google’s own infrastructure and shows you the rendered HTML, the screenshot and the resources that could not be loaded.

It is the best method available — and it only works for your own property. You cannot inspect somebody else’s page this way, which is a problem precisely when you need it most: when the page misusing your brand belongs to someone else.

3. Google’s own testing tools

The Rich Results Test accepts any public URL and fetches it from Google’s infrastructure with Googlebot’s user agent. So does PageSpeed Insights, which additionally returns a screenshot of what it was shown.

These are not the real crawler, but the request leaves Google’s address space rather than yours. A cloaker filtering by network sees a Google address and hands over the version it reserves for Google — and now you are holding it.

4. A request from Google’s own network

The same idea, one step further: a small script running on Google’s cloud fetches the page for you. The page sees a Google address and a Google signature, not yours. Together with the tools above, this covers the cloaking that keys on where the request comes from, not on what it calls itself.

This is what the checker at the top of this page does, alongside the plain visit and the visit that arrives from a Google search — and then it compares what each one was shown.

The one thing nobody can do

You cannot make a request from the real Googlebot’s IP addresses. Only Google can. Anything that claims otherwise is either using the user agent (method 1) or borrowing a Google address that is not the crawler’s (methods 3 and 4).

The reverse is possible: when a visit claims to be Googlebot, you can verify it by reverse DNS — a genuine one resolves to googlebot.com or google.com and back to the same address. That tells you whether a visitor was Google. It does not let you become one.

So content served only to the crawler’s own addresses cannot be fetched by anyone outside Google. It still leaves a trace: Google’s search result quotes text that is not on the page you can see. That is a probability, not proof — and it is why the checker compares the page with Google’s own snippet as well.

Why a page shows different things to different visitors

Not every difference is cloaking. Worth separating:

• The user agent. The crudest switch, and the easiest to catch.
• The referrer. The page behaves one way for someone who arrives from a Google search and another for someone who types the address. A common pattern: the search visitor is sent on to somewhere else entirely.
• The address the request comes from. Country, network, whether it looks residential. This is why our own check makes its plain visit from a residential address in the United States, where Googlebot crawls from.
• The visit number. Money page on the first visit, ordinary page afterwards — a cookie is enough.
• Rendering. Content assembled by JavaScript that the crawler does not run, or files it is not allowed to load. This is not deception, it is a mistake, and it is worth fixing rather than reporting.

Reading the difference

Two versions in front of you are not yet a verdict. What actually matters:

• Was a page served at all. A refusal is not a substitution. If one side gets 403 and the other 404, nobody was shown a page, and there is nothing to compare — the difference is between two refusals.
• One side a page, the other a wall. That is the strong signal: a visitor is stopped and a request from Google walks through, or the reverse.
• How much the text differs. Compared against how much the page differs from itself between two identical visits — plenty of pages rotate banners and ad slots on their own.
• Where you end up. A different final address for one of the visitors is worth more than any amount of differing text.

If the page is using your brand

A single check tells you about one page on one day. The pages that matter — the ones ranking on your name in markets you sell in — appear, disappear and come back, and the cloaking on them is switched on and off with them.

Serpward watches the results for your brand on a schedule, captures each unauthorised page it finds, runs this same check on it, and files the takedown with the evidence attached.

Find out what ranks under your name