Get Started in Minutes.
Authenticate with your API key. POST article URL or raw text. Get back bias spectrum, propaganda index, logical fallacies, and 15 diagnostic scores in JSON.
curl -X POST \
https://api.rhetoricaudit.com/v1/analyze \
-H "Authorization: Bearer YOUR_KEY" \
-d '{
"url": "https://example.com/article",
"model": "v20"
}' | jq
# Response:
{
"article_id": "abc123",
"bias_spectrum": {
"lean": "center-right",
"confidence": 0.87
},
"propaganda_index": 0.62,
"fallacies": [
{
"type": "Appeal to Authority",
"quote": "...",
"severity": "high"
}
],
"emotional_appeals": {
"fear": 0.71,
"hope": 0.42,
"urgency": 0.85
}
}Supported Formats:
- ✓ Article URLs (auto-extract text)
- ✓ Raw text (up to 10K characters)
- ✓ HTML content
- ✓ Social media posts
- ✓ Speeches and transcripts
Response Includes:
- ✓ Bias spectrum (left-center-right)
- ✓ Propaganda index (0-1)
- ✓ 24-type fallacy detection
- ✓ Emotional appeal scores
- ✓ Strategic silence analysis
- ✓ Source credibility signals
Built for Scale.
Newsroom Monitoring
Track competitor coverage across 100+ sources. Detect framing patterns and bias shifts in real-time.
Disinformation Detection
Identify propaganda campaigns and coordinated framing. Monitor for information operation patterns.
Research at Scale
Analyze 10,000s of articles for bias patterns. Build datasets for media studies and NLP research.
Content Moderation
Flag manipulative content automatically. Reduce manual review overhead with ML-backed scoring.
Recommendation Engines
Bias-score articles before recommending. Surface credible alternatives to sensationalism.
Political Intelligence
Track narrative shifts across parties. Detect coordinated messaging and propaganda intensity.
Scale from Prototype to
Millions of Scans.
API FAQ
What authentication method does the API use?
Bearer token authentication via Authorization header. Generate keys in your dashboard. Rotate keys anytime.
Can I batch multiple articles in one request?
Yes. Batch endpoint accepts up to 100 articles per request. Charged per article. Async job returns results in 30–60 seconds.
What's the SLA for the free tier?
Free tier: best-effort. Pay-as-you-go: 99.5% uptime SLA. Enterprise: 99.99% with dedicated support.
Can I get historical data for an article I scanned before?
Yes. Scan lookups by article ID are free. Re-analyzing cached scans is free. Only new scans count against quotas.
What models are available?
FME V20 (current, 93% accuracy), V19.1 (legacy, 100% band accuracy), V18.5 (historical). Specify via model parameter.
Do you have SDKs for common languages?
Python, JavaScript/Node, Go. See SDK docs. Community SDKs for Ruby, PHP, Java.