EdgeOutlier

EdgeOutlier is a web app with a lightweight edge agent that detects equipment and sensor anomalies locally, then syncs only high-signal events to the cloud. It’s built for teams who can’t afford to stream every datapoint or wait for centralized dashboards to notice problems. The agent runs simple, explainable models (seasonality + change-point + multivariate correlation checks) and produces a compact “anomaly packet”: what changed, when it started, likely contributing signals, and confidence. The web console lets you group assets, set baselines per shift/line, and route alerts to Slack/PagerDuty with suppression to avoid alert fatigue. It also supports offline buffering for spotty connectivity and a “known events” calendar (maintenance, product changeovers) to reduce false positives. This is not magic AI; it’s pragmatic detection tuned for noisy industrial data and constrained networks.