Rhetoric Audit

Forensic media analyzer. Detect bias, propaganda, 24 fallacies, strategic omissions in any article — 2-5 seconds.

PhD-level forensic media analyzer. Detect bias, propaganda, logical fallacies & strategic omissions — 13 metrics in seconds. FULL DESCRIPTION: Rhetoric Audit is a forensic media evaluation tool that analyzes any article you're reading — in real time — across 13 diagnostic metrics. It doesn't just tell you what an article says. It tells you how it's engineered to make you think. WHAT RHETORIC AUDIT ANALYZES: • Bias Spectrum Mapping — Places content on a precision gradient from Far Left to Far Right (not just binary labels) • Propaganda Index — Quantifies emotional loading, manipulation risk, and ideological engineering (0–100%) • Logical Fallacy Detection — Identifies 24 fallacy types (Ad Hominem, Strawman, False Dichotomy, and more) with exact quoted text • Strategic Silence — Reveals what the article deliberately doesn't say: missing counter-evidence, omitted perspectives • Emotional Resonance Triad — Separate Fear / Hope / Urgency scores (0–100 each) • Intent Transparency — Is the author informing or obfuscating? (0 = Obfuscating → 100 = Informing) • Evidence Validity Index — Verifiable data vs. anecdotal or assumed claims • Narrative Archetype — Hero Focus, Scapegoat Focus, or Victim Focus • Intellectual Depth (DOK) — Webb's Depth of Knowledge framework (DOK-1 through DOK-4) • Source Reliability — Author credibility scoring based on citation and epistemological humility • Philosophical Frame — The underlying ideological lens of the piece • Volatility Correlation — Societal impact sensitivity (LOW / MOD / HIGH) • PhD Analytical Deduction — A 3–4 sentence forensic synthesis of the complete rhetorical strategy WHO IT'S FOR: — Researchers & academics: Pre-coded rhetorical analysis, save hours in literature reviews — Journalists & fact-checkers: Verify narratives, not just facts — Policy analysts: Track media framing on policy topics — Critical readers: Read any article with forensic clarity HOW IT WORKS: 1. Install the extension (free, no API key required) 2. Sign up to receive 10 free analysis credits 3. Browse to any article and click the Rhetoric Audit icon 4. Receive a full forensic report in 2–5 seconds TECHNICAL: — Manifest V3 compliant — No inline scripts or eval() — Content Security Policy enabled — CORS-validated backend — Rate-limited API — Supabase Row-Level Security — you can only access your own analyses ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ WHAT'S NEW IN v2.1.4 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ v2.1.4 — Forensic Media Evaluation (FME) Framework This is a major upgrade from the previous release. Key changes: NEW — FME Framework (13 diagnostic metrics, up from 3): • Propaganda Index with 0–100% manipulation risk scoring • Logical Fallacy Detection across 24 types with exact quoted text • Emotional Resonance Triad: Fear / Hope / Urgency (independent scores) • Strategic Silence: surfaces deliberately omitted context • Intent Transparency: Obfuscating (0) → Informing (100) spectrum • Evidence Validity Index: data-backed vs. anecdotal claims • Narrative Archetype: Hero / Scapegoat / Victim framing detection • Intellectual Depth (Webb's DOK 1–4 framework) • Source Reliability score • Volatility Correlation (LOW / MOD / HIGH) • PhD Analytical Deduction (3–4 sentence forensic synthesis) NEW — Professional 2-Column Forensic Interface: • 820px redesigned popup with dedicated panels for each metric group • Expandable Logic Fractures: click any fallacy to see exact quote + impact • Contextual tooltips on every metric with plain-English explanations • Color-coded risk visualization (red = risk, green = proof, purple = emotional) NEW — Supabase Account & Credit System: • Sign up for a free account to save and revisit all analyses • 10 free credits on signup, no credit card required • 1 credit per article analysis • Full analysis history available at rhetoricaudit.com/dashboard NEW — Shareable Analysis URLs: • Every analysis generates a unique URL you can share • All 13 metrics are encoded in the URL — no login required to view IMPROVED — AI Engine: • 18+ prompt engineering iterations tested across 9 AI models • Primary: MiMo-V2-Flash (fastest, lowest cost) • Fallback: Qwen 2.5 72B Instruct • Chain-of-Thought processing for analytical consistency • Temperature 0.3 — prioritizes analytical accuracy over creativity IMPROVED — Security: • Content Security Policy (CSP) enabled • No eval() or inline scripts • Safe JSON parsing throughout • Rate limiting on analysis API • Input validation (minimum 100 characters required)