Data model
Understand content objects, key points, linked entities, clusters, and monitors.
Most endpoints return a content_item. The fields you usually render are at the top level, and the exact shape varies slightly by media_type.
Content
A content item represents one public item. Today that is usually a news article, wire item, or podcast episode.
Common top-level fields include:
| Field | Meaning |
|---|---|
id / content_id | Stable identifier for fetching the item later. |
media_type | news, wire, or podcast. |
published_date | When the source says the item was published. |
title | Display title when available. |
canonical_url | Best external URL for the item. |
source | Source name and domain. |
creator_name | Person or show-level creator when available. |
logline | One-sentence summary. |
informative_abstract | Longer structured summary. |
key_points | Concise extracted points that are useful for scanning, display, or alerts. |
linked_entities | Entities linked from the item. |
News and wire items may include fields such as article_id, language, author, and image. Podcast items may include fields such as episode_id, podcast_id, and segments when you fetch the item directly.
Full news article bodies and podcast transcripts are intentionally not returned by the public API yet.
Entities
Entities are stable records for people, companies, products, places, organizations, and other named concepts. Entity search helps you find an entity ID. Entity timelines return content where that entity appears.
News clusters
A news cluster is a story-level grouping of articles. The list endpoint returns non-stub clusters. Looking up a specific cluster ID can still return a stub when that is the stored object.
Monitors
A monitor is a saved search plus a schedule. Monitor results return the same content shape as search, and monitor search accepts the same parameters as content search.
That symmetry is the point: build one search form, then let users save it as a monitor without translating it into a second model.