What data is included with each article?#

Every APITube article arrives as one JSON object that bundles far more than the headline. Alongside the readable content, each article carries eight groups of machine-readable data: content and identity, classification (categories, topics, industries), entities with knowledge-graph links, three levels of sentiment, a source profile, readability and length metrics, media and summary, and engagement plus status flags. You do not run any of this yourself — the enrichment ships already computed, so you can filter, score, and route articles the moment you receive them.

This page is a decision-level map of what data you get per article. For the exact JSON shape field by field, see what’s inside a news article.

What content and identity fields does each article include?#

The readable core of every article:

• title — the headline, and description — a short plain-text summary with HTML removed.
• body — the full article text as plain text, and body_html — the same content with the original HTML markup preserved. You choose the form that suits NLP, indexing, or rendering.
• href — the canonical URL of the original article, and image — its lead image URL.
• published_at — the publish time in ISO 8601 (UTC).
• language — the detected language as an ISO 639-1 code, such as en.
• author — an { id, name } object when a byline was detected.

What classification data does APITube attach?#

Every article is tagged with thematic labels, all as arrays of objects (not bare strings), so each tag carries context:

• categories — IPTC Media Topics, each with an id, name, a relevance score, its taxonomy, and a self link to that category’s feed.
• topics and industries — extra thematic tags, each with an id, name, and a self link (topics also carry a score).
• keywords — a flat array of the salient terms extracted from the article.

Because scores and links come attached, you can rank a match by relevance or jump straight to every article in the same category.

What entity and knowledge-graph data is included?#

The entities array lists the real-world things the article mentions — people, organizations, places, and more. Each entity is one of nine types: person, location, organization, brand, product, natural-disaster, disease, event, or sport. For each entity you get:

• an id, name, and type;
• links to that entity’s APITube feed and, where known, its Wikipedia and Wikidata pages;
• a frequency count of how often it appears;
• a sentiment block scored toward that entity specifically, with a mentions breakdown of positive, neutral, and negative references;
• title.pos and body.pos arrays marking where the entity appears in the text;
• a metadata block for known organizations (details such as headquarters and key people).

A convenience array, locations_mentioned, distils the location entities that have a known country into { name, country, lat, lng } so you can map an article without walking the full entity list. For the full breakdown, see understanding entities.

What sentiment data comes with each article?#

APITube ships sentiment at three levels. The article-level sentiment object has three views — overall, title, and body — each with a numeric score from −1 (negative) to 1 (positive) and a polarity label of positive, negative, or neutral. On top of that, every entity carries its own sentiment, so you can tell whether a company was framed favourably even inside an article that reads neutral overall. See how sentiment analysis works for the scoring detail.

What source, quality and readability data is included?#

Each article is fully attributed and scored for quality:

• source — a publisher profile: domain, home_page_url, type, a political-leaning bias label, rankings.opr (the source’s OPR authority score), a location with country name and code, and a favicon URL.
• readability — reading-difficulty metrics including flesch_reading_ease, flesch_kincaid_grade, automated_readability_index, a difficulty_level, a target_audience, an estimated reading_age, plus average words-per-sentence and syllables-per-word.
• Length counters — words_count, sentences_count, paragraphs_count, characters_count, and read_time.

What media, engagement and status data is included?#

Finally, each article describes its attachments, reach, and state:

• media and links — arrays of { url, type } for attached media and outbound links; summary is an array of key sentences, each with its own sentiment.
• shares — social-share counts (total, facebook, twitter, reddit). These ramp up over roughly the first two hours after publication, so counts on a brand-new article are partial by design.
• story — the cluster this article belongs to, with an id and a uri to fetch every article covering the same event.
• Status flags — is_breaking marks high-importance news, is_duplicate flags likely near-identical reprints (an approximate signal, not exhaustive reprint detection), and is_free reports whether the original article is freely readable at its source (this is about the source, not your subscription).

Common Questions#

Is the data the same on every endpoint?#

Yes. Whether an article comes from a search, a single-article lookup, or the real-time stream, it is built into the same fixed object shape with the same field names. Fields that do not apply to a given article are still present but empty (an empty array, null, or 0), so your parser never has to guess whether a field will exist.

Do I have to receive all of this data?#

No. Use the fl parameter to request only the fields you need — for example fl=id,title,source.domain,sentiment.overall — which shrinks the response and speeds up parsing. The field names you pass are exactly the ones on this page. See how to select specific response fields.

Why is some data shortened on the free plan?#

On the Free plan and with test keys, long fields such as body, description, href, image, and source URLs are truncated and tagged with a marker like ...[Upgrade subscription plan]. The schema is identical to a paid response — the field names and structure never change, only the values are shortened. A paid plan with a live key returns the full content.