# Boolean search: raw queries and LinkedIn and X filters (https://snitchfeed.com/docs/listeners/boolean-search)



Advanced search gives you two extra tools for a keyword term: a **raw boolean query** you can type by hand, and **search settings** that filter LinkedIn and X results by author or engagement.

You do not need either to get started. If you are new to terms, read [Keywords](/docs/listeners/keywords) first. It covers the main term, AND, OR and skip words, and platforms.

The examples use the same scenario as [Keywords](/docs/listeners/keywords): you sell a **Notion competitor** and want posts where people complain about Notion or ask for alternatives.

## Write a raw boolean query [#write-a-raw-boolean-query]

Open a term, expand &#x2A;*Advanced Settings (boolean)** and find **Raw Boolean Query**. It shows the whole term as one line of text.

You can edit the text directly. When it is valid, the builder fields above it update to match. Changes in the builder fields show up in the text too.

### Syntax [#syntax]

| Syntax  | Meaning                   | Example                      |
| ------- | ------------------------- | ---------------------------- |
| `AND`   | Both sides must appear    | `notion AND export`          |
| `OR`    | Either side can appear    | `alternative OR replacement` |
| `NOT`   | Drop posts with this word | `notion NOT hiring`          |
| `"..."` | Exact phrase              | `"social listening"`         |
| `( )`   | Group words together      | `(SaaS OR startup)`          |

`AND`, `OR` and `NOT` work in any capitalization. Curly braces are not supported. Use `OR` instead.

### Example [#example]

A term for people looking to switch, without job posts or template sellers:

```text
notion AND (alternative OR replacement OR "switching from") NOT (hiring OR "template for sale")
```

In the builder, this is the main term `notion`, three words under **Must mention at least one of**, and two skip words.

### How words are sent to the search [#how-words-are-sent-to-the-search]

On X, LinkedIn and Hacker News, each term runs as a search on that platform.

* A word set to **Exact** is sent as a quoted phrase.
* If a skip list has more than 5 words, only 5 are sent to the search. Every skip word is still checked before a mention is stored.

## Filter LinkedIn results [#filter-linkedin-results]

Each term has a **LinkedIn Settings** card below the skip words. It only applies when LinkedIn is on for that term.

| Setting               | What it does                                                             | Limit              |
| --------------------- | ------------------------------------------------------------------------ | ------------------ |
| **Author job title**  | Only posts from authors with similar job titles                          | Up to 3 titles     |
| **Author company**    | Only posts from authors at these companies (paste LinkedIn company URLs) | Up to 4 companies  |
| **Author industries** | Only posts from authors in these industries                              | Up to 5 industries |
| **Post type**         | **Any**, **Person** or **Company**                                       | One choice         |

Keep these two rules in mind:

* **Each job title is its own search.** Three titles means three LinkedIn searches per scan, and each is billed. See [Credits](/docs/credits).
* **LinkedIn allows at most 5 operators.** A single query can hold at most 5 AND, OR and NOT operators. The editor warns you when a term goes over.

## Filter X results [#filter-x-results]

Each term has an **X (Twitter) Settings** card below the skip words. It only applies when X is on for that term.

| Setting           | What it does                 |
| ----------------- | ---------------------------- |
| **Minimum likes** | Skips posts with fewer likes |
| **Language**      | Only posts in this language  |

## Next steps [#next-steps]

* [Keywords](/docs/listeners/keywords): set up terms and pick what to track.
* [Scans & backfill](/docs/listeners/scan-intervals): how often terms are checked.
* [Credits](/docs/credits): what each scan costs.
