> ## Documentation Index
> Fetch the complete documentation index at: https://docs.brightdata.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Business Search query syntax

> Business Search structured query reference: 7 operators, and, or, not, equals, in, range and text, plus nesting and text matching rules.

A Business Search structured query is a nested JSON object sent as the `query` property in Ludicrous mode. Field operators create conditions, and logical operators combine them.

## Which operators does a structured query accept?

| Operator | Meaning                                           |
| -------- | ------------------------------------------------- |
| `and`    | All child conditions must match                   |
| `or`     | At least one child condition must match           |
| `not`    | Exclude records matching its child condition      |
| `equals` | Match an exact field value                        |
| `in`     | Match any value in a supplied list                |
| `range`  | Match numeric bounds using `>`, `>=`, `<` or `<=` |
| `text`   | Search a supported text field                     |

`and`, `or` and `not` take conditions as children. `equals`, `in`, `range` and `text` take a field name and a value.

## How do you combine conditions?

Nest logical operators to build the boolean expression you need. This company query searches for software or technology companies in the US or GB, excluding records whose organization type is `Educational`:

```json theme={null}
{
  "and": [
    {
      "or": [
        { "text": { "industry": "software" } },
        { "text": { "industry": "technology" } }
      ]
    },
    { "in": { "headquarters_country_code": ["US", "GB"] } },
    { "not": { "equals": { "organization_type": "Educational" } } }
  ]
}
```

Use `in` instead of chaining `or` plus `equals` on the same field. It suits country lists, industry lists and any enumerated value set.

## How does text matching work?

A `text` condition can contain a string, as in the company example above, or an object with `value` and `mode`. Use `mode: "all"` when every search word must match. The words do not need to form an exact phrase:

```json theme={null}
{
  "text": {
    "current_title": {
      "value": "machine learning engineer",
      "mode": "all"
    }
  }
}
```

`mode` accepts seven values, and the API rejects any other value with an HTTP 400 that lists them:

| Mode                   | Behavior                                                                  |
| ---------------------- | ------------------------------------------------------------------------- |
| `top-bm25`             | Default. BM25-ranked partial matches                                      |
| `all`                  | Every search word must match                                              |
| `all-or-top-bm25`      | All-word match first, backfilled with BM25 matches when there are too few |
| `top-idf`              | IDF-ranked partial matches                                                |
| `all-or-top-idf`       | All-word match first, backfilled with IDF matches                         |
| `any`, `top-bm25-daat` | Accepted, not documented here                                             |

`top-k` caps how many candidate records the text condition retrieves, 100 by default. No mode matches an exact phrase.

`text` works only on text fields. On a typed field such as `headquarters_country_code` it returns HTTP 400 with `Operator does not support field type 'string'. Supported types: array<text>, text`.

`equals` does not work on text fields. The operator a field accepts depends on its type: fields typed as text, such as `industry` and `current_title`, accept only `text`, and sending `equals` to one returns HTTP 400 with `Operator does not support field type 'text'`. `equals`, `in` and `range` are for typed fields such as `headquarters_country_code`, `city` and `followers`.

## How do you express a numeric bound?

`range` takes a field name and one or two comparison keys. Supply both keys for a closed interval:

```json theme={null}
{
  "range": {
    "employees_in_linkedin": { ">=": 50, "<=": 500 }
  }
}
```

Supply one key for an open bound, such as `{ "followers": { ">=": 1000 } }`.

## Which fields can you search?

Field names differ by category, and the operator a field accepts depends on its type. The full lists are on [Company search](/products/business-search/company-search#which-company-field-names-should-you-use) and [People search](/products/business-search/people-search#which-people-field-names-should-you-use). A field that appears in `view` output is not necessarily searchable: people `connections` is returned in `summary` but cannot be used in a `range` condition.

When a condition is rejected, correct the request rather than retrying the same body. Business Search returns HTTP 400 with a message naming the field, the operator and the types the operator supports. See [Business Search error handling](/products/business-search/error-codes).

## Frequently asked questions

### Does the query object change between company and people searches?

The operators are identical. The field names are not. The company category uses `industry`, `headquarters_country_code` and `linkedin_followers`. The people category uses `current_title`, `country_code` and `followers`. See [Company search](/products/business-search/company-search) and [People search](/products/business-search/people-search).

### Can I use a structured query with Instant or Smart mode?

No. `ludicrous` takes a structured object, and `instant` and `smart` take a natural-language string. A mismatch is rejected with HTTP 400: an object in Instant returns `request: "query" must be a string`, and a string in Ludicrous returns `request: "query" must be of type object`. See [What is Business Search?](/products/business-search/introduction#what-does-each-mode-do).

### Does adding a field to view filter the results?

No. `view` selects output fields only. Search conditions live in `query`, and the two are evaluated independently.