ScamWarning by MCPCIO

ScamWarning detects scams, phishing, and fraud on every page you visit — right in your browser toolbar. HOW IT WORKS ScamWarning uses a three-stage assessment to evaluate each page: 1. Local pattern matching runs entirely in your browser. Eight categories of scam indicators are checked without any data leaving your device. 2. Threat database lookup checks the page against known threats and verified safe sites using semantic search. 3. AI analysis (only for pages not already in our database) uses Claude by Anthropic to assess novel threats. Results are stored anonymously so future visitors benefit instantly. WHAT IT DETECTS - Phishing and credential harvesting - Financial fraud and investment scams - Tech support scams and fake warnings - Government and brand impersonation - Urgency manipulation and pressure tactics - Misleading advertisements and clickbait - Suspicious e-commerce and fake stores - Data harvesting and privacy violations - Malware distribution sites THREE RISK LEVELS Green = Safe. No threats detected. Yellow = Moderate. Some suspicious patterns found. Proceed with caution. Red = Threat. Significant scam or fraud indicators detected. TWO PROTECTION MODES Click-to-Scan (default): Click the toolbar icon to scan the current page. Nothing is sent until you click. Always-On: Automatically scans every page you visit. You choose. MULTILINGUAL ScamWarning responds in the language of the page you're viewing. Spanish pages get Spanish assessments. Japanese pages get Japanese assessments. PRIVACY FIRST - No browsing history collected - No personal information stored - No IP address tracking - No analytics or telemetry - No ads, no premium tier, no upsells - No account required - Anonymous install ID for rate limiting only (never linked to your identity) - Full source code available on GitHub BUILT BY MCPCIO ScamWarning is built by MCPCIO (mcpcio.com), a privacy-first AI platform. We built this because we believe basic online safety should be free for everyone. Open source under MIT License.