Norm Mcp

This server gives you statistical baseline checking for detecting anomalies in your data. Feed it values with category labels, and it tells you whether they're within normal ranges or flag as outliers based on the distribution it's learned. Useful when you're building monitoring dashboards, fraud detection pipelines, or any system where you need to know if a metric has suddenly gone sideways. It works locally over stdio, so there's no external API dependency. The implementation is straightforward: you send categorical data points, it builds statistical models per category, and queries return whether new values are anomalous. Reaches for basic z-score or IQR methods rather than complex ML models, which means fast responses and easy to reason about.