See the bias in your news.

Your AI watchdog for objectivity.

How It Works — Live Example

Claude: The headline uses mostly neutral language with only mild positive connotation in 'gesture,' presents the event factually without strong editorial judgment, and reports a verifiable action, though 'Holy Week gesture' adds slight interpretive framing by characterizing the motivation. | OpenAI: The headline uses neutral language, presents the pardoning as a factual event without overt judgment, and is largely verifiable, though the term 'gesture' may imply a positive connotation.

🟣Claude
Loaded Language82
Framing85
Factuality90
🟢OpenAI
Loaded Language90
Framing80
Factuality95
LLM Agreement: 94%Final score is the average of both AI engines
Sign in to vote for the better analysis· 1 vote so far

Overall Rankings — Last 30 Days

Since Mar 29, 2026
#Channel
1
2
3
4
5
6
7
8
9
10
NPR
NPRUS
75
11
12
13
14
15
16
17
18
19
20
21
22
23

How scores work: Each article is scored 0–100. Headline language, framing, and factuality account for 75% of the score; lead image analysis (emotional manipulation, sensationalism) accounts for 25%. On topic pages, each channel also shows its directional lean — the perspective its headlines most frequently favor (e.g. “pro-Palestinian”, “pro-government”). A ⚠ Coverage Bias alert appears when a channel both over-covers a topic (2×+ the peer average) and scores below 40/100. Learn more →

Disclaimer: All scores and labels are generated by AI and may be inaccurate. Analysis quality improves over time as more data is collected and models are refined. Scores should be used as a general guide, not as definitive judgments of journalistic quality.