A Complete List of Boolean Search Operators, With Syntax and Examples

What is the full list of boolean search operators?

The eight boolean search operators supported in some form by every query language are AND, OR, NOT, quotation marks, parentheses, wildcards, proximity operators and field prefixes. Each one changes how the engine combines the terms you typed rather than what those terms mean, so the same words in a different structure return a different set. Learn the eight as a group, because a query that fails almost always fails at the join, not at the vocabulary.

Wildcard, proximity and field prefix syntax varies by engine, while quotation marks, AND, OR, NOT and parentheses hold almost everywhere. Build the skeleton out of those five.

How does operator precedence change a query's results?

Precedence is the order an engine resolves operators when you mix them, and the standard order is NOT first, then AND, then OR. That order is invisible in the query string, which is why a query with no parentheses returns a set unrelated to what you meant. Wrap every OR group in parentheses as a habit, even when the order looks like it works out.

These two queries look equivalent and are not:

reddit OR bluesky AND monitoring

(reddit OR bluesky) AND monitoring

The first resolves as reddit OR (bluesky AND monitoring) because AND binds tighter than OR, so it returns every record mentioning reddit for any reason, plus the small set where bluesky and monitoring appear together. The second requires monitoring in every record. No engine warns you which one you wrote.

The same trap sits inside a NOT clause. alternative OR review NOT hiring removes hiring from the review half only, so every hiring thread that says alternative survives. Write (alternative OR review) NOT hiring and the exclusion covers both.

What do wildcards and proximity operators do?

A wildcard replaces a variable ending on a stem, and a proximity operator sets the maximum distance between two terms. Both trade precision for reach, so add them after the exact phrase version proves too narrow, one at a time, re reading the results each time.

A wildcard on invoic* is safe because the stem carries meaning. A wildcard on pro* catches product, project, promotion and profit, and turns a clean query into a firehose. Proximity fails the other way: a window of 3 misses ordinary phrasing, and a window of 25 stops meaning anything in a long post.

What do field prefixes do in a search query?

A field prefix restricts a term to one part of a record, such as the title, the author, the community or the date. A keyword in a thread title signals what the discussion is about, while the same keyword buried in comment 300 signals an aside. Use a title prefix when you want the conversation to be about your term, and drop it for every passing mention.

Prefix names and quirks change per platform. Our page on X advanced search operators covers that platform's field syntax in full, and the Bluesky search operators page covers the shorter from, to, since, until, lang and domain set Bluesky documents.

Why does a listening query return noise or miss the conversation?

A listening query fails in one of two directions: too broad and the feed fills with records nobody reads, too narrow and the conversation you needed never appears. Both feeds look functional, because a broad query shows plenty of results and a narrow one shows a clean handful. Measure the query instead of trusting how it looks.

The measurement is precision, the share of returned results that are relevant:

Precision = relevant results / total results reviewed x 100

Read the first 50 results and tag each one. If 12 are relevant, precision is 12 / 50 x 100 = 24 percent, and a reviewer is discarding four records for every one worth reading. Below 50 percent, people stop opening the feed within two weeks. The trade against recall, the share of real conversations the query catches, is the whole design problem, and precision versus recall works through where to sit on it.

The broad failure comes from a generic term or an ambiguous brand name, and the fix is a required context term plus exclusions built from the false positives you just read. Those exclusions are negative keywords, and they carry more weight than the positive terms do. The narrow failure comes from an exact phrase where buyers use three phrasings, and the fix is an OR group of the words they actually type, taken from real threads.

How do you build and refine a boolean query?

Build the query in four clauses, then refine it against real data until precision holds. The loop matters more than the first draft, because nobody writes a working query before seeing what it returns. Budget an hour for the first pass and 15 minutes a quarter after.

The four clauses run in order: a topic group of quoted phrases joined by OR, an intent group joined by OR and attached with AND, an exclusion clause joined with NOT, and any field prefixes last.

("invoicing software" OR "billing tool" OR "invoice app")
AND (recommend OR alternative OR "looking for" OR "switching from")
NOT (hiring OR salary OR giveaway OR coupon)

Then check the other direction. Find one real thread the query should have caught, search for it by hand, and confirm the query returns it. If it does not, the topic group is missing a phrasing your buyers use. RedReplier takes a website and keywords, surfaces mentions ranked by buying intent, and explains with AI why each one was flagged, so the refine loop becomes reading the explanations you disagree with. See Reddit keyword monitoring for that setup.

Query design sits inside the wider practice described in what social listening is and what it tells you, and the same operators drive a brand monitoring setup where the topic group is your own name and its misspellings.

Frequently asked questions

Do boolean search operators have to be uppercase?

Most engines require AND, OR and NOT in uppercase and read a lowercase "and" as an ordinary word to search for. A few accept either case. Write them uppercase everywhere, because a query that works lowercase on one platform searches for the word itself on the next.

What is the difference between NOT and the minus sign?

They do the same job in most engines: both remove records containing the term that follows. The minus sign attaches to its term with no space, as in -hiring, while NOT sits between two expressions. Some engines support one form only, so check the platform's documentation first.

How many terms can a boolean query hold?

The engine sets the hard cap, and platforms publish their own character or clause limits. The practical limit comes first: a query nobody on the team can read is a query nobody maintains. Split a long one into two named queries by intent instead of stacking 40 OR terms.

Why does my query return results that contain none of my keywords?

Check the parentheses first. An unparenthesized OR next to an AND resolves as a OR (b AND c), which returns everything matching a alone, and those records need no connection to the rest of the query. Wrapping the OR group fixes it.

Should I use a wildcard or list the word forms with OR?

List the forms with OR when there are three or fewer, and use a wildcard when the stem has many endings. The OR list is explicit, so you can read the query later and know what it catches. A wildcard on a stem under five characters pulls in unrelated words.

Do boolean operators work the same on every platform?

The five core operators, quotation marks, AND, OR, NOT and parentheses, behave the same almost everywhere. Wildcards, proximity operators and field prefixes are where dialects diverge, and some platforms support none of the three. Test each query on the platform that will run it.

How do I know if my boolean query needs fixing?

Read the first 50 results and tag each one as relevant or not. That precision score tells you where the query stands: a search returning 12 relevant results out of 50 sits at 24 percent, well below the 50 percent line where a feed stays worth opening.

Why do people stop checking a search feed?

A feed that scores under 50 percent precision loses its audience within two weeks, because a reviewer keeps discarding four records for every one worth reading. The fix isn't to check more often, it's to raise the precision before anyone opens the feed again.

Which boolean operators are worth learning first?

Quotation marks, AND, OR, NOT and parentheses work almost the same on every platform, so they form the skeleton of any query. Wildcards, proximity operators and field prefixes vary by engine, so learn the five core operators before touching syntax that changes from platform to platform.

How much time should a boolean query take to build?

Budget about an hour for the first draft, since a working query rarely survives contact with real data on the first try. After that, plan on roughly 15 minutes a quarter to re run the loop and keep precision where it needs to be.