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

# Response fields

> Every field of a search response and of an article.

## Search response

**query** `string`

The query as received, with control characters replaced and the ends trimmed.

**effort** `string`

The effort that ran: `low`, `medium` or `high`.

**sort\_by** `string`

The order of the results: `relevancy` (best match first), `published_at` or
`crawled_at` (newest first).

**took\_ms** `integer`

How long the search took on our side, in milliseconds.

**window** `object`

The time window that was searched.

**properties**

**search\_window** `string`

Length of the window as a `search_window` value (`7d`, `2h`, `15m`): the one you sent, your plan's limit when it was cut, or `7d` by default.

**hours** `number`

The same length in hours, a fraction for windows in minutes.

**from** `string`

Start of the window: `to` minus its length (UTC, ISO 8601).

**to** `string`

End of the window: when the search ran (UTC, ISO 8601).

**clamped** `boolean`

True when the window you asked for was longer than your plan allows and was cut. Absent otherwise.

**count** `integer`

Number of results returned, at most `limit`.

**results** `object[]`

The articles, in the order of `sort_by`. See below.

**excluded** `string[]`

The terms of the `NOT` clauses of `q`. Absent when `q` has no `NOT`.

## Article

**id** `string`

Article id.

**title** `string`

Headline as published.

**title\_en** `string`

English translation of the headline. Non-English articles only.

**summary** `string`

One-line summary in the article's language. Absent when the article has none.

**summary\_en** `string`

One-line summary in English. Non-English articles only.

**snippet** `string`

Start of the publisher's description, up to your plan's snippet length (200
characters on the free plan).

**url** `string`

Link to the article.

**source** `object`

The publisher: `domain` (without `www.`) and `name`.

**language** `string`

ISO 639-1 code. Chinese is `cn` (simplified) or `tw` (traditional).

**country** `string`

Country of the publisher, ISO 3166-1 alpha-2.

**published\_at** `string`

When the article was published (UTC, ISO 8601).

**crawled\_at** `string`

When NewsCatcher crawled it (UTC, ISO 8601).

**content\_type** `string`

The type of content the article was classified as, for example `news_report`.

**story** `object`

The [story](/docs/news/stories.md) it belongs to: `id` and `newsrooms`.

**score** `number`

Ranking score. Only the order of the results means something, not the value.

## Enterprise fields

Enterprise plans can add heavy fields to every result with `include`. Other
plans get `403` with `enterprise_field`. Your plan's fields are in
`extra_fields` of `GET /subscription`. For everything else Enterprise plans
add, see [Enterprise plans](/docs/get-started/enterprise-plans.md).

| `include` | Field | What it holds |
| - | - | - |
| `description` | `description` | The description the publisher wrote for the article. |
| `entities` | `entities` | People (`ner_PER`), organizations (`ner_ORG`), locations (`ner_LOC`) and other names (`ner_MISC`), plus the same found in the English translation. |
| `sentiment` | `sentiment` | Sentiment of the headline and of the article, from -1 (very negative) to 1 (very positive). |
| `themes` | `themes` | Themes of the article. |
| `translation` | `translation` | English headline and summary of a non-English article. |
| `embedding` | `embedding` | A 1,024-value vector of the article, rounded to 4 decimal places. |


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.