Google Search Operators for SEO Audits 2026

Administrator in Search Engine Optimization Sep 24 / 3 weeks ago

It is 8:40 on a Monday morning in Columbus, Ohio. The head of ecommerce at an outdoor gear store on Shopify has just typed site: and her domain into Google. The first page is not her hero products. It is a wall of collection URLs with sort and filter parameters, plus a PDF size chart from 2021. Her note lands in your inbox at 8:41: "Is Google indexing the wrong half of our site?"

You can answer that in ten minutes with Google search operators. You can also get it badly wrong in the same ten minutes, because the number under that query is a guess, not a count. I have watched smart people paste a site: total into a board deck as if it were the index. I did it myself, early on, and the figure changed when I hit refresh.

This guide shows you how to use operators the way a triage nurse uses a thermometer. Fast, cheap, and always followed by a real test.

Here is the short version. Search operators are a fast way to ask the index a question. They are not a way to measure the index. Google's own site: page says the list of URLs it returns is not always exhaustive, and that bigger sites should not expect to see all their URLs. So the first rule is simple: use Google search operators to find leads, then confirm every lead in Search Console.

Four angles here that most cheat sheets skip:

• Which operators Google documents today (only a handful), and which are folklore that still happens to work.
• The Search Console cross-check that turns a hunch into a finding you can put in an audit.
• Why the same query returns different pages in Denver and Munich, and what the EEA changes did to that.
• What broke when cache: and related: went away, and why the num=100 change in September 2025 matters to anyone paging through site: results.

The objection I hear most is that crawlers make operators pointless. Screaming Frog and Sitebulb show you what your site serves. Operators show you what Google chose to keep. Those are different questions, and a good audit asks both. Scope: Google web search, with a short Bing detour, and no dorking on sites you do not own. Everything here fits inside the first hour of a technical SEO audit built around a short, ranked list of fixes.

What are Google search operators and which ones does Google still document?

Google search operators are short commands you add to a query, such as site:, filetype:, or a minus sign, to narrow what Google returns. On the pages Google last updated on 2025-12-10, Search Central documents just four for site debugging: site:, filetype:, imagesize:, and src:. The Search Help page adds quotes, the minus sign, before:, and after:. Everything else is undocumented.

I read all three primary pages on 2026-09-24. The Search Central debugging page opens by saying Google "supports several search operators" and lists the four above. inurl:, intitle:, allintitle:, intext:, OR, the wildcard, AROUND(n), and define: appear on none of those pages. Practitioners report that all of them still work in 2026. Google has just never promised them, so treat each one like a macro in a colleague's spreadsheet. Useful, until the day it is not.

I built the table below from the three Google pages read on 2026-09-24. The documented column comes straight from those pages. The undocumented rows rest on wide practitioner use, not on any Google promise.

Why is site: a triage tool and not a census?

The site: operator shows a sample of the URLs Google holds under a domain or path, not the full set. Google says so on the operator's own page: the list is not always exhaustive, and bigger sites should not expect to see all their URLs. So use it to spot what looks wrong, and count with Search Console instead.

The site: operator page on Search Central also warns that site:https://www.example.com and site:https://example.com/ do not return the same results. The host and the exact prefix matter. Two more limits bite in practice. Google stops paging long before the full set, and practitioners commonly report a ceiling of a few hundred results for a site: query. Since the num=100 parameter died in September 2025 you also page ten at a time. A thermometer tells you there is a fever. It never tells you which infection. That is the whole job of site: in an audit.

How do you find what Google has indexed for a site in ten minutes?

Run five queries in order: the bare site: query, site: with inurl:? for parameters, site: minus your www host to expose subdomains, site: scoped to one section, and site: with filetype:pdf. Scan the first three pages of each. Write every surprise down as a lead, not a finding, and stop at ten minutes.

1. site:example.com (2 minutes). Skim three pages and note anything you would not want a customer to land on.
2. site:example.com inurl:? (2 minutes). Sort, filter, and session parameters here mean faceted navigation is leaking filter URLs into the index.
3. site:example.com -site:www.example.com (2 minutes). Staging, dev, and old regional hosts show up here.
4. site:example.com/blog/ (2 minutes). Compare the shape of results with what your CMS says it published.
5. site:example.com filetype:pdf (2 minutes). Old price lists and size charts live here.

Pages that appear here but sit nowhere in your navigation are orphans. The crawl depth guide for large sites covers how to find and reconnect them.

How do you cross-check operator results against Search Console?

For every lead, open Search Console. The Page indexing report gives you the real indexed and not indexed counts. URL Inspection gives a verdict for one URL, with the canonical Google picked. The Sitemaps report shows how many submitted URLs made it in. When Search Console says indexed and site: does not show the page, trust Search Console.

Google's own site: page points you to URL Inspection when a URL is missing from results. Budget 15 minutes for ten URLs. The practitioner's guide to Search Console walks through the Page indexing statuses, and the XML sitemaps guide explains why the submitted versus indexed gap in the Sitemaps report is the count that matters. The most common failure: an auditor subtracts a site: total from the Page indexing total and reports the gap as a problem. The gap is noise. Report only URLs you inspected.

How do you spot keyword cannibalization with intitle: and inurl:?

Search site:example.com intitle:"your phrase" to list your own pages that carry that phrase in the title. Two or more results aimed at the same intent are a lead. Confirm in the Search Console Performance report by filtering on the query, which shows which URLs earned impressions for it and whether Google keeps swapping them.

Take a French publisher with a guide titled "assurance auto" and a comparison titled "meilleure assurance auto". The intitle: query lists both in seconds. Google search operators cannot tell you which one Google prefers, or whether the swap costs clicks. Search Console can. The cannibalization guide covers the merge, redirect, or differentiate decision, and the canonical tags guide covers the case where the two URLs are near duplicates rather than rivals. Add inurl: when the phrase lives in the slug, not the title. Budget 20 minutes for ten keywords.

How do you find leaked PDFs, staging copies, and gated content?

Run site:example.com filetype:pdf for documents, site:example.com -site:www.example.com for stray hosts, and a quoted sentence from behind your paywall to test whether Google is serving the full text. Each hit is a lead, not a verdict. The fix is usually a noindex header or a canonical, and almost never a robots.txt Disallow.

A Munich news publisher with a metered paywall can paste one sentence from a locked article in quotes and see in seconds whether a syndication partner or a staging host is exposing it. The paywalled content guide explains how to stay indexed without cloaking. For a PDF you want gone, remember that Disallow stops Googlebot from reading the page, so it never sees your noindex. The robots.txt and noindex guide lays out the X-Robots-Tag route. Budget 10 minutes.

How do you check a migration cleanup with Google search operators?

After a domain or path migration, run site:olddomain.com weekly and expect it to shrink over weeks, not days. Run site:newdomain.com inurl:old-path to catch URLs that should have redirected. Spot-check every surviving URL in URL Inspection, because old URLs can linger in site: results long after Google has processed the 301 redirect.

A London agency that moves a client from a .co.uk to a .com domain can still see the old host in site: results for months. That is the index catching up, not a failure. Panic sets in when someone treats the lingering count as proof the redirects broke. The redirects and site migration playbook gives the launch-day and 90-day monitoring schedule. Operators add one thing crawlers cannot: a view of what Google still remembers. Budget 20 minutes a week for the first quarter.

How do you use before: and after: to find stale content?

Search site:example.com/blog/ before:2023 to list indexed posts Google last updated before 2023. The Help page frames these operators around last updated, not first published. Dates accept a plain year, slashes, or dashes. Treat the output as a shortlist for refresh or pruning, then rank it by clicks in Search Console.

The Refine Google searches Help page documents before: and after: with examples like before:2025/11/01 and after:2019-03-01. Google's date is its own estimate, so a post you edited last month can still show as old. The content refresh and pruning guide covers what to do with the list. Budget 15 minutes for a blog with a few hundred posts.

How do you audit a competitor with Google search operators?

Start with a site:competitor.com query to see the rough shape of its indexed footprint, then site:competitor.com inurl:blog for publishing shape and site:competitor.com intitle:"keyword" for topic coverage. Use quotes to find your brand mentioned on their pages. Then move to Ahrefs or Semrush for the numbers, because operators show presence and never traffic.

A US home goods retailer can learn in five minutes that a rival indexes a comparison page for every product family. That is a lead for a content gap, not a verdict. The competitor analysis playbook shows how to turn it into a ranked task list. One honest limit: link: died years ago, so operators cannot show backlinks. Use the Search Console Links report, Ahrefs, or Majestic, as the backlink audit guide explains.

Which Google search operators stopped working, and what replaces them?

The cache link left results in early 2024, and cache: itself stopped working in September 2024, when its doc page vanished. related: left the Search Central docs in July 2023. link:, info:, the plus sign, and the tilde died years earlier. The num=100 URL parameter stopped working in September 2025. Replacements: the Wayback Machine, competitor reports in Ahrefs or Semrush, and the Search Console Links report.

Google's SearchLiaison account said on X in July 2023 that related: "hasn't really worked that well for some time" and was "rarely used", as Search Engine Roundtable reported on 2023-07-19. Search Engine Journal reported on 2024-09-24 that the cache: docs were gone and that Google had added Wayback Machine links to its About this result panel. Write-ups from Locomotive in September 2025 and Botify in October 2025 tied a sudden drop in impressions, sharpest on desktop, to rank trackers that could no longer load 100 results per page. If your chart shows a cliff near 2025-09-12, read the traffic drop decision tree before you blame an update.

Do the same operators return different results in the USA and Europe?

Yes. Google localizes results by country and language, so a site: query run from Denver and from Munich can surface different pages in a different order. The operators behave the same. What changes is the mix, and in the EEA the page layout itself changed under the Digital Markets Act. Check both markets when a site serves both.

Add gl=de and hl=de to the Google URL, or use a VPN, to see the German view. Since 2026-09-08 the EEA layout shows an aggregator unit for hotel, flight, train, bus, and product queries, with local business queries added on 2026-09-18, which pushes direct sellers down without changing what site: returns. The DMA search changes guide shows how to measure that with a Search Console country split. For hreflang sites, run site: per country host and compare.

Do these operators work on Bing, and what about Naver?

Bing documents its own set on Microsoft's Advanced search keywords page, including site:, filetype:, intitle:, inbody:, and ip:. Most Google habits carry over, with different limits. Naver, which splits the South Korean market with Google, uses its own syntax, so a Korean audit needs a different toolkit. Never assume a Google trick works elsewhere.

Bing also gives you something Google does not: the AI Performance report in Bing Webmaster Tools, which shows the pages Copilot cited. The Bing SEO and Copilot guide covers that and IndexNow. Bing's site: results carry their own limits, so treat them as a second sample rather than a tiebreaker. Budget 10 minutes to repeat your five triage queries there.

Back to that Monday in Columbus. The right reply is not a number. It is a list: filter URLs in the index, a stale PDF, and a staging host, each with a URL Inspection screenshot next to it. Google search operators found the leads in ten minutes. Search Console made them findings.

My priority order for your next audit: run the five triage queries, inspect every surprise, fix leaks with noindex rather than Disallow, then run the intitle: pass for overlap. Do the competitor pass last. It is fun, and fun is how audits run long.

A prediction: Google will keep trimming operators quietly, documented or not, the way it trimmed cache: and related:. Anything you rely on that is not on the Search Central page should have a Search Console fallback written next to it today. So here is my question for you: which operator would hurt most if it vanished tomorrow, and what would you use instead?

Frequently asked questions about Google search operators

Does the site: operator show all the pages Google has indexed?

No. Google's own site: operator page, last updated 2025-12-10, says the list of URLs is not always exhaustive and that bigger sites should not expect to see all their URLs. In practice you get a sample, by most practitioner accounts a few hundred results at most, and the order shifts between refreshes. Use it to spot surprises such as parameter URLs, stray subdomains, or old PDFs. For a real count, open the Page indexing report in Search Console, and for one URL use URL Inspection. That is the record. The operator is the flashlight.

Why does the site: count change every time I search?

Because it is an estimate, not a count. This question fills Reddit threads every month, and the answer never changes. Google has said for years that the figure is a rough estimate, it is plainly rounded, and the sample shifts with data center, location, and time of day. Two people running the same query in Chicago and Paris can see different totals within a minute. Do not put the number in a report. If you need a trend, use the Page indexing chart in Search Console, which reads the index itself.

Is Google dorking the same as using search operators, and is it legal?

Google dorking is the security world's name for stacking operators to find exposed files, login pages, or config data. The syntax is the same site:, inurl:, and filetype: you use for SEO. Running those queries is generally lawful in the USA and Europe, because you are only reading public search results. Acting on what you find against a site you do not own can cross into unauthorized access. Keep audits to sites you control or have permission to review, and report leaks on client sites rather than testing them.

Does the cache: operator still work in 2026?

No. Google's SearchLiaison confirmed in early February 2024 that the cache link was gone from results, and the cache: operator itself died around 2024-09-24, when Search Engine Journal reported its documentation page gone too. Google now links to the Internet Archive's Wayback Machine from the About this result panel, a change it made in September 2024. For audit work, the Wayback Machine shows what a page looked like on a date, and URL Inspection in Search Console shows the HTML Google last crawled. Between those two you lose nothing cache: gave you, and you gain a crawl date.

What is the difference between intitle: and allintitle:?

intitle: applies to the single word or quoted phrase right after the colon, so site:example.com intitle:pricing finds titles with "pricing" while the rest of the query matches anywhere. allintitle: applies to every word after it, so allintitle:pricing plans 2026 needs all three in the title. Neither appears on a Google documentation page today, so both are undocumented, and allintitle: has a reputation among practitioners for returning odd samples on large sites. For cannibalization checks, use intitle: with quotes and confirm in the Search Console Performance report.

Are Google search operators case sensitive?

The operator names are not. site:, SITE:, and Site: are treated as the same operator, and so are filetype: and inurl:. Your search terms are not case sensitive either. The one strict thing is spacing: Google's Help page says not to put a space between the operator and your term, so site: example.com breaks and site:example.com works. OR must be uppercase to act as a logical operator rather than a plain word. When a query returns nothing, check spacing and quotes before you assume the operator died.

Why does Google show a captcha when I run many operator searches?

Because a burst of operator queries from one IP looks like a scraper, and Google throttles scrapers. Industry write-ups tied the num=100 shutdown in September 2025 to bulk scraping, and many practitioners report stricter throttling since. Slow down, use a normal browser, and spread the triage pass over a few minutes. Queries from a shared office network or a Cloudflare WARP exit can trigger it faster. If you need hundreds of checks, stop using operators and export the URL list from Search Console or a crawler instead.

How do I find pages Google indexed that are not in my sitemap?

Start with site:example.com and skim for URLs your CMS never meant to publish, such as the tag archives WordPress creates by default, filter combinations, or print views. Then open the Page indexing report in Search Console and filter by the status for indexed pages not submitted in a sitemap. That status is the authoritative list, and it exports. Operators help you guess where the leak is, for example inurl:tag or inurl:?sort. Search Console tells you how big it is. Fix the source with noindex or canonical tags first.

Can search operators show me my backlinks?

Not anymore. link: was retired years ago and no documented operator replaced it. You can find mentions of your brand with quotes, and you can find pages that embed one of your images with src: in Google Images, but neither is a link report. Use the Links report in Search Console for what Google sees, then Ahrefs, Semrush, Majestic, or Moz for a bigger sample and anchor text. For a toxic link review, operators are the wrong tool from the first minute.

Which tools replace search operators for a full index audit?

Search Console first, always, because it reads the index directly. Then a crawler: Screaming Frog costs 279 US dollars a year on its pricing page as of 2026-09-24 and crawls 500 URLs free, while Sitebulb writes the friendlier report. For large sites, JetOctopus, Botify, or Lumar join the crawl to server logs. Operators sit in front of all of these as the ten-minute smoke test. They are free, they need no access, and they show you what Google kept, which no crawler can.