CogBias Detector — Cognitive Bias & Manipulation Finder

Detects cognitive manipulation, propaganda techniques, and logical fallacies on web pages and in YouTube videos using AI.

CogBias Detector is an AI-powered Chrome extension that analyzes web pages and YouTube videos for cognitive manipulation techniques, propaganda, and logical fallacies in real time. Whether you're reading the news, scrolling social media, or watching YouTube — this tool helps you see through manipulation and think critically. 🔍 WHAT IT DETECTS (14 techniques): • Loaded Language — emotionally charged words designed to bypass rational thinking • Appeal to Fear — creating a sense of threat to suppress critical thinking • False Dichotomy — presenting a situation as only two extreme choices • Ad Hominem — attacking a person instead of addressing their argument • Cherry Picking — selectively citing only convenient facts • False Consensus — "everybody knows…", "it's been proven…" • Whataboutism — deflecting criticism by pointing fingers elsewhere • Concept Substitution — subtly replacing one term with a misleading one • Appeal to Authority — citing experts irrelevant to the topic • Bandwagon Effect — "everyone does it" social pressure • Slippery Slope — unfounded chain of catastrophic consequences • Straw Man — distorting an opponent's position to easily defeat it • Emotional Blackmail — manipulating through guilt, shame, or obligation • Hasty Generalization — drawing broad conclusions from isolated cases 📖 WEB PAGE ANALYSIS: • Analyze the entire page with one click • Or select specific text and analyze only that • Manipulative passages are highlighted directly on the page • Click any highlight to see: technique name, severity level, confidence score, and a human-readable explanation • Three sensitivity levels — Strict (only obvious manipulation), Balanced, or Maximum (flags everything suspicious) 🎬 YOUTUBE VIDEO ANALYSIS: • Paste any YouTube video URL, or the extension auto-detects YouTube pages • Automatically extracts subtitles (manual or auto-generated) • Analyzes the transcript for manipulation techniques • Results displayed with timestamps [MM:SS — MM:SS] • Click any result to jump directly to that moment in the video • Works with subtitles in multiple languages (Russian, English, and more) ⚙️ HOW IT WORKS: 1. Click the extension icon on any web page or YouTube video 2. Choose your mode: Full Page / Selected Text / YouTube 3. Set sensitivity level (Strict / Balanced / Maximum) 4. Click "Analyze" — results appear in seconds 5. Highlights appear on the page; click them for details The extension uses a large language model (LLM) to analyze text against known cognitive bias and propaganda patterns. Each detected issue includes the exact quote, technique name, severity rating, confidence score, and an explanation of why it may be manipulative. 🔒 PRIVACY & SECURITY: • Nothing happens until YOU click "Analyze" — no background scanning • Only the visible text of the page you choose to analyze is sent to the server • No browsing history, personal data, cookies, or tracking • Analysis results are cached locally in your browser and cleared on close • Open source: https://github.com/AlexB0nch/cogbias-detector 💡 USE CASES: • Fact-checkers and journalists verifying media content • Students learning about logical fallacies and propaganda • Researchers studying media manipulation techniques • Anyone who wants to read the news more critically • Teachers demonstrating cognitive biases in class 🌐 LANGUAGES: The extension interface is available in English and Russian. Analysis works with content in any language supported by the AI model. 📬 FEEDBACK: Found a bug or have a suggestion? Open an issue on GitHub: https://github.com/AlexB0nch/cogbias-detector/issues