CRED-1 – Open domain credibility dataset for on-device pre-bunking

CRED-1 is an open-domain dataset designed to train on-device AI models for pre-bunking misinformation by providing labeled examples of credible and non-credible claims. It targets developers and researchers building lightweight, privacy-preserving fact-checking tools for mobile or edge devices. The project is interesting because it bridges the gap between large-scale credibility research and practical, offline deployment for real-time misinformation prevention.