InTruth

Real-time fact-checking of political speech and debates

🚀 Beta out now ! No API key needed, uses multiple LLMS, still available in 16 languages ! ! https://intruth-beta.vercel.app/ 1.2.6: added source bias classifiers from the MBFC dataset; includes (extreme/center) left/right as well as neutral from 3000+ known international sources! InTruth is for anyone who wants to stay informed during any video, but particularly, live debates, news interviews, press conferences, and political events. Now works for multiple languages like Spanish, Hindi, Portuguese, and German! This extension requires users to provide their own AI API key. Transcripts are sent directly to the AI provider selected by the user for fact-check analysis. The extension developer does not receive, store, or sell transcript data or API credentials. This extension listens to live audio from videos and broadcasts, identifies factual claims as they are made, and provides instant verdicts based on authoritative sources. Instead of waiting for post-event analysis, you can see claim evaluations as participants speak. Features: • Live Fact-Checking: Detects and evaluates check-worthy claims in real time as speakers talk. • Speaker Attribution: Tracks who made each claim + attributes claims to the correct participant. • Evidence-Based Verdicts: Classifies claims as True, Substantially True, False, Misleading, or Unverifiable using reputable sources. • Political Debate Coverage: Designed for debates, interviews, town halls, press conferences, and campaign events. • Works Alongside Video Content: View verdicts without interrupting the live stream or video you're watching. How to use InTruth: - Open a supported video or livestream. - Start the extension. - The extension transcribes speech in real time; choose which participant is speaking when prompted. - Check-worthy claims are identified in real time and auto-evaluated against authoritative information sources. - Verdicts and explanations appear as the discussion unfolds. Notes: This extension focuses on factual, verifiable claims. Opinions, predictions, campaign promises, rhetorical statements, and subjective judgments are generally not fact-checked. While the system strives for accuracy, AI-generated fact checks may occasionally be incomplete or incorrect. Users should review cited evidence and use independent judgment when evaluating important claims. Perfect for voters, journalists, researchers, students, and anyone who wants additional context during live political and public affairs coverage!! Super excited to share :-) -Risha P