NoiseCut

NoiseCut is a web app (with optional desktop agent) that sits next to your SIEM and continuously measures alert quality: duplicates, low-signal rules, noisy sources, and recurring false positives. It ingests alert metadata via SIEM APIs (not raw logs), then builds a “noise map” by rule, source, host group, and time window. Teams can simulate suppression, thresholding, and routing changes before pushing them, with rollback and change tracking. It also generates weekly “top 20 fixes” recommendations and a lightweight KPI dashboard (MTTA, alert volume per analyst, false-positive rate proxies). This is a combination traditional + AI app: AI helps cluster similar alerts and summarize root causes, but every change requires human approval and is expressed as explicit, auditable rule/routing edits. The goal is pragmatic: reduce alert fatigue without claiming to replace your SOC.