Check this page with an assistantOpens a chat asking it to summarise this article and name the evidence behind each claim.
Claude opens with the prompt on your clipboard: Anthropic does not support prefilled prompts on the web, and we would rather copy it than ship a button that drops it.
The Working Set
You do not need forty operators. Eight cover essentially every practical use, and combining them is where the value is. The rest are either niche, unreliable, or gone.
| Operator | What it does | Genuinely useful for |
|---|---|---|
site: | Restricts results to one domain or path | Seeing what kinds of pages are indexed |
"exact phrase" | Matches the phrase exactly | Finding copies of your own text |
-term | Excludes results containing a term | Removing noise from a diagnostic search |
intitle: | Requires a word in the page title | Finding pages built around a specific topic |
inurl: | Requires a string in the URL | Locating templated sections and parameters |
filetype: | Restricts to a file format | Finding indexed PDFs you forgot about |
OR | Matches either term | Covering variant spellings in one search |
* | Wildcard within a phrase | Finding phrasings around a fixed structure |
Two syntax rules cause most failures. No space after the colon, so site:example.com works and site: example.com does not. And the operators themselves are case-insensitive, so the uppercase versions you sometimes see are stylistic rather than required.
The Ones That Are Gone
Google retires operators quietly and guides keep listing them for years afterwards. Two are worth knowing about specifically, because they appear in advice published this year and neither has worked for a long time.
cache:was removed in early 2024. It showed Google's stored copy of a page and was a genuinely useful diagnostic for what the crawler had seen. The replacement for that job is the URL Inspection tool in Search Console, which shows the rendered HTML Google actually fetched and is more accurate than the cache view ever was.
link: was deprecated in 2017. It never returned a complete backlink list even when it worked, which is worth remembering when someone quotes it as evidence about a competitor. daterange: still exists in various forms and returns incomplete results, so treat anything it tells you as a partial view rather than an answer.
Diagnostics That Take Seconds
Operators earn their place as fast checks you would otherwise need a tool for. These are the ones worth committing to memory, because each answers a question that comes up constantly.
- Is this specific page indexed?
Search
site:plus the exact URL. No result strongly suggests it is not indexed, which is a completely different problem from ranking badly and needs the sequence in indexing diagnostics. - What is indexed that should not be?
site:your domain and page through the results. Tag archives, parameter URLs, staging subdomains, and old PDFs surface here, which is the fastest way to spot the index bloat described in duplicate content. - Has my site been used to publish spam?
site:your domain combined with terms you would never sell, or simply reading deep into the results for pages you do not recognise. This is the first check in hacked site recovery. - Who mentions us without linking?
Search your brand in quotes with
-site:your own domain to exclude yourself. The result is a list of unlinked mentions, which is the easiest link source in how to get backlinks. - Which roundups should I be on?
intitle:best, combined with your category, finds the list pages that exist. Cross-reference against what assistants actually cite before pitching, per earning third-party citations. - Has my content been copied?
Put a distinctive sentence in quotation marks. If it appears elsewhere, you have found scraping, which matters most when the copy is outranking your original.
What site: Does Not Tell You
The result count is an estimate and a volatile one. People check it twice in a week, see a large difference, and conclude something has gone wrong. Use it qualitatively, to see what kinds of pages are indexed, and use Search Console for anything where the number matters.
It also does not show everything. Operator searches run against a filtered view and are not a complete listing, so absence is suggestive rather than conclusive. The URL Inspection tool gives a definitive answer for a specific URL, which is why it is the right instrument when the question is important.
The reasonable division of labour: operators for exploration, where you do not yet know what you are looking for, and Search Console for verification, where you need to be sure. The routine that uses the second properly is in the Search Console operator's guide.
Using Them on Competitors
Operators work on any domain, which makes them a free competitive research tool. What they show you is structure and coverage, which is frequently more useful than the metrics people pay for.
site: a competitor with inurl: reveals how they template sections, whether they run location or comparison pages, and how large those programmes are. site: combined with intitle: and a topic shows how deeply they cover it, which is a direct read on the coverage argument in topical authority.
What it cannot show is what works. A competitor with four hundred location pages may be getting nothing from them, and the operator count says nothing about traffic or revenue. Treat it as a map of what they built rather than evidence of what succeeded, which is the same caution that applies throughout competitor analysis.
Questions About Search Operators
- Do Google search operators still work in 2026?
Around twenty-five of them work reliably, including the ones that matter most for diagnostics: site:, intitle:, inurl:, filetype:, exact-match quotation marks, OR, and the minus exclusion. A handful return incomplete results and a few well-known ones have been retired entirely.
- Which operators have been removed?
cache: was removed in early 2024 and link: was deprecated back in 2017, yet both still appear in guides published this year. daterange: has become unreliable rather than formally removed. Google can change or silently break any operator without notice, which is why testing before relying on one matters.
- Is site: an accurate count of my indexed pages?
No, and treating it as one causes real confusion. The number shown is an estimate and fluctuates considerably. Use it to see what kinds of pages are indexed, which is genuinely useful, and use Search Console for anything where the actual count matters.
- Are search operators case sensitive?
The operators themselves are not, so site: and SITE: behave identically. What does matter is the syntax: no space after the colon, and quotation marks around phrases you want matched exactly. Most operator failures are a stray space rather than a deprecated command.
- What are operators actually most useful for?
Fast diagnostics you cannot get elsewhere without a tool. Checking whether a specific page is indexed, finding pages you did not know existed on your own domain, spotting hacked content, locating unlinked brand mentions, and finding the roundup pages worth pitching for inclusion.
