BiasLens - AI Bias & Hallucination Detector

BiasLens: Real-Time Rhetoric & AI Hallucination Detector Navigate the modern web with clarity. BiasLens is a lightweight, real-time forensic reading assistant that inspects suspect excerpts, exposing emotional framing, logical fallacies, and unsubstantiated AI hallucinations directly within your browser. Whether reading breaking news, product reviews, scientific write-ups, or social commentary, BiasLens equips you with transparent, objective insight into the language on your screen. KEY FEATURES - Decoupled Scoring Engine: Separates rhetorical slant from empirical verifiability. * Bias Severity Score (0–100): Measures emotional manipulation, loaded framing, and logical fallacies. * Hallucination Risk (0–100): Flags unverified empirical assertions, fabricated consensus, and pseudoscientific bluffing. - Context-Aware Evaluation: Never evaluates text in isolation. BiasLens reviews your selected excerpt alongside surrounding paragraphs and document headings to interpret tone and intent accurately. - Granular Phrase Extraction: Isolates the exact offending substrings in the text that triggered the alert, making rhetorical tactics visible and educational. - Violation Categorisation: Pinpoints the core manipulation tactic at play, including Ad Hominem, Authoritative Bluffing, Emotional Framing, Loaded Language, Sensationalism, and False Equivalency. - Zero Configuration: Ready out of the box with no complex setup, API keys, or configurations required. HOW IT WORKS 1. Highlight any suspect or heavily phrased text on any webpage. 2. Click the floating BiasLens badge or right-click and choose "Analyse with BiasLens". 3. Instantly review the diagnostic card containing the primary violation, dual scores, a concise explanation, and the highlighted keywords. PRIVACY-FIRST DESIGN - Zero Browsing History Logged: BiasLens only processes the specific text excerpt you highlight and transmit for evaluation. - No Tracking or Profiles: We do not collect cookies, browsing telemetry, or personal identity markers. - Transparent & Open Source: BiasLens is fully auditable. Inspect the repository at https://github.com/mistovek016/BiasLens.git