What is an entity extraction API?

An entity extraction API detects and labels the named things in text — people, organizations, locations, and more. The APITube entity recognition API returns these on every article as structured data your models and agents can act on.

What entity types does APITube recognize?

APITube recognizes entities across 9 types: person, location, organization, brand, product, event, disease, natural disaster, and sport. Each detected entity is returned with its type, so you can filter to exactly the kind you care about.

Are entities linked to a knowledge graph?

Yes. Every entity resolves to Wikipedia and Wikidata links, so an agent can disambiguate the right OpenAI, the right city, or the right product instead of guessing from the surface text.

Does each entity have its own sentiment?

Yes. Alongside the link and type, each entity carries its own sentiment score from −1 to +1, so a single article can be positive about one company and negative about another.

Can I search or filter news by entity?

Yes. You can search and filter articles by entity, and each entity is returned with its position in the text, so you can highlight or extract mentions precisely.

How do I find an entity to filter by?

Use the suggest tool to look up an entity by name and get its identifier, then filter articles by that entity. The suggest tool is also exposed on the hosted MCP server for AI agents.

Is entity extraction available on the free plan?

Yes. Entities, their knowledge-graph links, and per-entity sentiment are part of the enriched schema returned on every plan, including the free tier.

Trusted by 6,000+ developers

Entity Extraction for News

Last updated: July 22, 2026

Turn articles into structured entities. APITube tags people, organizations, brands, products, locations, and more on every article — each linked to a knowledge graph and scored for sentiment — so your model knows exactly who and what a story is about.

Realtime news articles from 177 countries

in 60 languages

Benefits

Every article, resolved into structured entities

An entity extraction API pulls the people, places, organizations, and things out of text. APITube tags entities across 9 types on every article — person, location, organization, brand, product, event, disease, natural disaster, and sport — each linked to Wikipedia and Wikidata with its own sentiment score, so an agent can disambiguate and act on the right entity.

Get Free API Key

Learn more about API

Entities across 9 types on every article

Each entity linked to Wikipedia and Wikidata

Per-entity sentiment score from −1 to +1

Entity position (start and end) in the text

Look up entities by name via the suggest tool

Search and filter articles by entity

59 languages, 177+ countries

Use cases

Discover the possibilities

Our solutions are designed to help you achieve your goals. Whether you are a small business or a large corporation, we have the right solution for you.

RAG

Disambiguation for RAG

Give a retrieval pipeline knowledge-graph links so it grounds on the right entity, not a namesake.

Monitoring

Company and people monitoring

Track every mention of a company, product, or person across 59 languages, resolved to one entity.

Graph

Knowledge-graph building

Populate a graph from news with typed, linked entities and their relationships over time.

Entity-based search

Let users filter coverage by a specific organization, brand, or event rather than raw keywords.

Features

Don’t waste time on complex features

Complex features made simple. Our API provides a simple way to access news articles from around the world. We provide a simple, consistent, and easy-to-use API to access news articles from thousands of sources.

News API

Export data in many formats
Industry Monitoring
Brand Monitoring
Market Intelligence
Risk Management
Competitive Intelligence
Media Monitoring
Sentiment Analysis
Trend Analysis
Story Grouping
Forecasting social trends
Multi-language Support
Audience Engagement
Geographical Analysis
Real-time Breaking News
Historical Data Access
Custom News Feeds
News Aggregation
Content Filtering
Over 50 integrations

Extract Additional Data

Industries
Locations
Persons
Organizations
Brands
Events
Disasters
Diseases
Links
Media
Images & Videos
Hashtags
Authors
Source
Duplicate Detection
Publisher Rank
Article Sentiment
Readability Score
Language Detection
and more

Analysis

Sentiment Analysis
Analysis of public opinion
Categorization
Financial Analysis
Trend Analysis
Story Grouping
Content Summarization
Entity Recognition
Keyword Extraction
Topic Modeling
Event Detection
Named Entity Recognition
Text Classification
Controversy Detection
Trust Score Analysis
Engagement Metrics
Source Bias Detection
Quality Ranking
Spam Detection
Emotion Detection

Advanced Searching

Search by Location
Search by Date Range
Search by Source
Search by Category
Search by Industry
Search by Sentiment
Search by Story
Search by Publisher Rank
Search by Language
Search by Entity
Search by Keywords
Boolean Search
Proximity Search
Faceted Search
Range Queries
Search by Author
Search by Media Type
Search by Breaking News
Search by Read Time
+ over 65 more filters

More than 500,000+ sources

APITube is trusted by teams around the world to help them build and deliver amazing digital experiences faster than ever before.

Last updated: 8s ago
Total sources: 512.645k
Requests yesterday: 25.768
Articles added yesterday: 137.054k
Total articles: 3.84b

Frequently asked questions

: An entity extraction API detects and labels the named things in text — people, organizations, locations, and more. The APITube entity recognition API returns these on every article as structured data your models and agents can act on.
: APITube recognizes entities across 9 types: person, location, organization, brand, product, event, disease, natural disaster, and sport. Each detected entity is returned with its type, so you can filter to exactly the kind you care about.
: Yes. Every entity resolves to Wikipedia and Wikidata links, so an agent can disambiguate the right OpenAI, the right city, or the right product instead of guessing from the surface text.
: Yes. Alongside the link and type, each entity carries its own sentiment score from −1 to +1, so a single article can be positive about one company and negative about another.
: Yes. You can search and filter articles by entity, and each entity is returned with its position in the text, so you can highlight or extract mentions precisely.
: Use the suggest tool to look up an entity by name and get its identifier, then filter articles by that entity. The suggest tool is also exposed on the hosted MCP server for AI agents.
: Yes. Entities, their knowledge-graph links, and per-entity sentiment are part of the enriched schema returned on every plan, including the free tier.