Google search operators: how to audit your SEO with search commands

Google search operators are the scalpel of SEO: one well-written query tells you in seconds what a site actually has indexed. Here are the operators Google documents today, the ones that still work undocumented, and the ones that are long dead even though half the internet keeps recommending them.

What Google search operators are and what they are for in SEO

An operator is a command you type inside the search box, not in Search Console. It filters results in ways a normal search cannot, with nothing installed. Google adds a rule almost nobody respects: there is no space between the operator and the term. site:example.com works; site: example.com does not.

In day-to-day SEO work they have three concrete uses:

• Verify: check what is indexed and what never made it in.
• Find: surface pages you did not know you had, leftovers from old versions and parameter noise.
• Read: see what competitors have indexed and who talks about your brand, with or without a link.

Where to type them and how to combine them

Everything happens in the search bar, and their real value lies in combining them: intitle:"keyword" site:example.com is a mini report in one line. Advanced Search does the same without memorising syntax, but it is slower for repeated audits. Quotes force an exact match ("exact phrase") and the hyphen excludes (-word): with it you strip your own site out and see the rest of the internet.

The operators Google officially documents

The official page, titled "Refine Google searches", lists only six techniques and presents them as "some of the most popular operators". It never promises a complete table, and that detail matters:

1. "word or phrase": exact match.
2. -word: exclude that term from the results.
3. site:example.com: results from that site or domain only.
4. before:2025/11/01 or before:2019: pages updated before that date.
5. after:2026/03/10 or after:2020: pages updated after that date.
6. filetype:pdf: a single file type.

The last two combine into a range: after:2019-03-01 before:2019-03-05. If a tutorial teaches you forty operators and the official documentation only recognises these six, the rest are not necessarily dead: they carry no guarantee and no support, and can change without notice.

The ones that work without being documented

There is a family of operators used daily in SEO that never appears on the official list: intitle: and allintitle: for title words, inurl: and allinurl: for the URL, intext: and allintext: for the body, AROUND(X) for two terms close together, the logical operator OR, parentheses for grouping, the wildcard * and numeric ranges with two dots.

The honest way to present them: they work in practice, but with no contract. Their behaviour is irregular, results come back incomplete and nothing obliges Google to keep them alive. Use them as a lead, never as definitive proof: if an intitle: returns zero results, confirm it with site: and URL Inspection before concluding your site is not indexed.

The ones that no longer work (and still appear in tutorials)

This is the reason the article exists. Several commands repeated in courses and blogs have been retired for years:

• cache: was removed. Google announced it in March 2024 and the documentation disappeared in September, as Search Engine Journal reported; the URL now redirects to a changelog. The alternative is the Internet Archive or URL Inspection.
• The plus sign (+) is gone: to force an exact word, use quotes.
• related: left the official documentation in July 2023 and still answered intermittently, as Search Engine Roundtable noted. Documentation and behaviour do not always move together.
• link: was retired years ago: to see links, use the Links report in Search Console or an external tool, as covered in the link building guide.
• info:, ~ and num= do not appear in the documentation and their effect is null or unpredictable.

The detail that sums it up: one of the most quoted lists in the industry, the 44-command list from Ahrefs, still describes cache: and related: as if they worked. In SEO, the list everyone copies is not the source; the official documentation is.

10 SEO audits with search operators

These ten queries are ready to paste as they are: swap in the domain and you have a quick review of any site.

1. How many of your pages Google sees. site:domain.com. The count is an estimate, not the real index figure; the reliable number lives in Search Console. If it looks absurd, the site is not indexed or you are on the wrong domain.
2. Indexed noise that should not be there. site:domain.com inurl:tag, inurl:?s=, inurl:/page/. Parameters and endless taxonomies competing and burning crawl budget; the detail is in meta robots and noindex and in faceted navigation.
3. Keyword cannibalisation. intitle:"keyword" site:domain.com and then "keyword" site:domain.com. If two of your URLs show up with similar intent, decide which one you keep; the method is in the cannibalisation guide.
4. Indexed files nobody reviews. filetype:pdf site:domain.com, plus filetype:doc and filetype:xls. An inventory of outdated dossiers and brochures; when HTML is the better call is covered in PDF and SEO.
5. Protocol and subdomain duplicates. Compare site:domain.com, site:www.domain.com and site:http://domain.com. If all three return URLs, live versions of the same content exist; review www, HTTPS and duplicate content.
6. Pages outside the sections you know. site:domain.com -inurl:/blog/ -inurl:/en/. Whatever is left is old landing pages, test folders or orphan pages.
7. Brand mentions without a link. intext:"brand" -site:domain.com. Who talks about you without linking, something the links report never shows; this is entity SEO reading.
8. Who copies your content. "a unique sentence from your article" -site:domain.com. Pick a long, unlikely sentence, never the title.
9. What competitors have indexed. site:competitor.com and site:competitor.com inurl:/category/. A map of their indexed structure to contrast with yours.
10. Leftovers from an old version. site:old-domain.com and site:domain.com after:2019 before:2021. Pages that should have died after a migration, and content untouched for years.

The limits: what operators do not tell you

• The numbers are personalised estimates: they shift with country, language and history. To compare, rerun the query in a private window, with the country fixed and Verbatim on; even then, two people can see different figures.
• site: is not the index: it does not show every indexed page, and Google has said so. URL Inspection gives the real status of one URL: indexed or not, chosen canonical, rendered version and last crawl. For the exact figure, Search Console.
• An undocumented operator returning zero results proves nothing: the absence could be syntax, personalisation or a command that no longer works. Verify it with a second source before writing a report.
• They do not replace your own crawl or log analysis: operators show what Google returns in a query, logs what Googlebot actually requests. Nor do they replace a technical SEO audit.
• There are no operators for what AI sees: if you want to know what Google extracts from a URL, the answer is not in the search bar.

Checklist: 5 queries to keep handy

1. site:your-domain.com: the general index picture, and the question "why is this URL missing".
2. site:your-domain.com inurl:blog: how much of each section you have indexed.
3. intitle:"keyword" site:your-domain.com: express cannibalisation check for a keyword you care about.
4. filetype:pdf site:your-domain.com: file inventory.
5. -site:your-domain.com "your brand": who mentions you without linking.

A final reminder: every finding gets confirmed in Search Console, or in your logs, before you touch anything.

Conclusion: verify before you opine

Three ideas. Operators turn a suspicion into a data point in ten seconds, and they are free. The list in circulation is half dead, so separate what is documented from what is not. And site: is a lead: the truth is in Search Console.

The next step is concrete: open a private window, paste site:your-domain.com, write the number down and compare it with the pages report in Search Console. That gap is already information about your site, and it fits any SEO analysis you run afterwards.