What is a media bias API?

A media bias API returns news articles filtered by the political leaning of the outlet that published them. APITube assigns each source it can classify a label of left, center or right, and exposes it as the source.bias parameter. That lets one request retrieve a topic as covered by a single side, or split a topic across all three.

Does APITube score the bias of the article or of the source?

Of the source. The label describes the publisher, not the individual piece of writing. A centrist outlet can still publish an opinion column, and a left-leaning outlet can still publish straight wire copy. Treat source.bias as a property of the publication and use sentiment fields if you need signal about the text itself.

How many articles carry a bias label?

More than 9.4 million: roughly 3.8 million from left-leaning sources, 1.9 million from centrist sources, and 3.8 million from right-leaning sources. Not every source in the index is classified, so a slice of any result set will have no label. Add facet.field=source.bias to a query to see the exact split for your own filters before you rely on it.

How do I get left, center and right coverage of the same story?

Run the same query three times, changing only source.bias. For example title=election&source.bias=left, then =center, then =right. Because every other filter stays identical, the difference between the three result sets is attributable to the leaning of the publishers.

Can I request more than one leaning at once?

Yes. source.bias accepts a comma-separated list, so source.bias=left,right returns both flanks and skips the centre. To go the other way and exclude a leaning, use ignore.source.bias with the label you want dropped.

Can I count how much coverage came from each side?

Yes. Add facet=1&facet.field=source.bias to any query and the response includes a breakdown by leaning alongside the articles. The facet returns numeric codes — 0 for left, 1 for center, 2 for right — plus a bucket for sources with no label, which is useful for reporting share of voice.

How do you decide which outlets are left, center or right?

Each source carries a single editorial-leaning label stored against the publication, applied at the level of the outlet rather than per article. Because any bias classification is a judgement call, we publish the label as data rather than as a verdict — combine it with source.rank.opr, which measures authority independently of politics, when you need a second signal.

What can I build with a bias filter?

Balanced news readers that show both flanks side by side, blindspot detectors that surface stories one side is not covering, media studies datasets with leaning as a variable, and newsroom tools that check whether a desk is over-indexed on one part of the spectrum.

Is the bias filter available on the free plan?

Yes. source.bias, ignore.source.bias, and bias facets are part of the standard filter set and work on every plan, including free. The free plan limits results per request, history depth, and adds a delay to fresh articles, but the filter itself is not restricted.

Trusted by 6,000+ developers

Media Bias API

Last updated: August 23, 2026

See the same story the way each side told it. APITube labels every source it can as left, center or right, so a single parameter splits your feed along the political spectrum — no scraping bias charts, no manual source lists.

Free Trial - No credit card required

Rated 5 stars by over 848 users

Realtime news articles from 177 countries

in 60 languages

Benefits

One parameter, three points of view

A media bias API returns news filtered by the political leaning of the publisher that produced it. APITube labels each source left, center or right and exposes that label as the source.bias filter, so one query can pull the same story as covered by each side. More than 9.4 million articles currently carry a bias label — 3.8M left, 1.9M center, 3.8M right.

Get Free API Key

Learn more about API

source.bias filter: left, center or right

9.4M+ labelled articles — 3.8M left, 1.9M center, 3.8M right

Combine several leanings in one call: source.bias=left,right

Exclude a leaning entirely with ignore.source.bias

Break any query down by leaning with facet.field=source.bias

Stack with source.rank.opr.min to keep only high-authority outlets

Works alongside sentiment, entities, and 59 languages

Full filter set on the free plan

Interactive Demo

Pick a scenario and see real API responses — filtered, structured, and ready to use. No credentials needed.

News API Scenarios

Ready to build with News API?

Discover recipes and code examples for common news data workflows.

Workflow examples

Full data models

API Response

✓ Tesla Stock Sentiment

Status: ok

Loading...

Key Features Demonstrated:

Sentiment AnalysisCompany ExtractionRich MetadataImage URLsSource Attribution

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.

Product

Balanced news readers

Show every story in three columns — left, center, right — so readers see the spread instead of one feed.

Product

Blindspot detection

Compare counts per leaning on the same query to surface stories one side is barely covering.

Research

Media studies datasets

Export a corpus with leaning as a variable alongside sentiment, entities and authority.

Newsroom

Source diversity audits

Check whether your own aggregation over-samples one part of the spectrum.

Comms

Coverage mapping

See which flank is driving a story about your organisation before you respond to it.

Balanced RAG context

Retrieve grounding passages from across the spectrum so a model is not fed one side only.

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 300,000+ sources

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

Last updated: 4s ago
Total sources: 300.000k
Requests yesterday: 20.888
Articles added yesterday: 157.157k
Total articles: 3.85b

Frequently asked questions

: A media bias API returns news articles filtered by the political leaning of the outlet that published them. APITube assigns each source it can classify a label of left, center or right, and exposes it as the source.bias parameter. That lets one request retrieve a topic as covered by a single side, or split a topic across all three.
: Of the source. The label describes the publisher, not the individual piece of writing. A centrist outlet can still publish an opinion column, and a left-leaning outlet can still publish straight wire copy. Treat source.bias as a property of the publication and use sentiment fields if you need signal about the text itself.
: More than 9.4 million: roughly 3.8 million from left-leaning sources, 1.9 million from centrist sources, and 3.8 million from right-leaning sources. Not every source in the index is classified, so a slice of any result set will have no label. Add facet.field=source.bias to a query to see the exact split for your own filters before you rely on it.
: Run the same query three times, changing only source.bias. For example title=election&source.bias=left, then =center, then =right. Because every other filter stays identical, the difference between the three result sets is attributable to the leaning of the publishers.
: Yes. source.bias accepts a comma-separated list, so source.bias=left,right returns both flanks and skips the centre. To go the other way and exclude a leaning, use ignore.source.bias with the label you want dropped.
: Yes. Add facet=1&facet.field=source.bias to any query and the response includes a breakdown by leaning alongside the articles. The facet returns numeric codes — 0 for left, 1 for center, 2 for right — plus a bucket for sources with no label, which is useful for reporting share of voice.
: Each source carries a single editorial-leaning label stored against the publication, applied at the level of the outlet rather than per article. Because any bias classification is a judgement call, we publish the label as data rather than as a verdict — combine it with source.rank.opr, which measures authority independently of politics, when you need a second signal.
: Balanced news readers that show both flanks side by side, blindspot detectors that surface stories one side is not covering, media studies datasets with leaning as a variable, and newsroom tools that check whether a desk is over-indexed on one part of the spectrum.
: Yes. source.bias, ignore.source.bias, and bias facets are part of the standard filter set and work on every plan, including free. The free plan limits results per request, history depth, and adds a delay to fresh articles, but the filter itself is not restricted.

Help Center

InsightsAugust 3, 2026

Resources

News API Documentation All Parameters Query Builder Sentiment Analysis Pricing Help Center

Developers

Documentation How It Works Structured News Data Interactive Demo

Trusted by 6,000+ developers

Media Bias API

One parameter, three points of view