OpenPMX gives small and mid-sized manufacturers the same AI-powered predictive maintenance capabilities that Fortune 500 companies pay $100,000/year for — completely free.
Built by a controls engineer who understands manufacturing — not just software.
Detects abnormal machine behavior the moment it starts — not after it becomes a crisis.
Predicts how many days before a component reaches its failure threshold.
Every machine gets a 0–100 health score. No technical knowledge needed to read it.
Critical alerts trigger automatically when a machine needs immediate attention.
Run entirely on your own network. Your sensor data never leaves your facility.
docker-compose up — that's it. Full platform running in minutes on any server.
From sensor data to health scores in under a second.
Your existing vibration, temperature, or pressure sensors send data to OpenPMX.
OpenPMX learns what "normal" looks like for your specific machines.
Every machine gets a real-time score from 0 to 100.
Schedule maintenance when it's convenient — not when machines break down.
Tested on the NASA IMS Bearing Dataset — 34 days of real bearing sensor data running until failure.
| Component | Final Health Score | Status | Outcome |
|---|---|---|---|
| Bearing 1 | 80 / 100 | Healthy | Correctly identified as healthy |
| Bearing 2 | 86 / 100 | Healthy | Correctly identified as healthy |
| Bearing 3 | 0 / 100 | Critical | ✓ Correctly identified failure — 908% vibration increase |
| Bearing 4 | 39 / 100 | Warning | Correctly flagged degradation weeks before failure |
Join manufacturers across the US using OpenPMX to prevent costly downtime.