Predictive-failure dashboards. Work order management at hospital scale. Offline-capable mobile CMMS. No license fee. No vendor lock-in.
The data that built
the case for open CMMS.
of unplanned downtime traces to six failure modes
Bearing wear, seal degradation, misalignment, lubrication failure, corrosion, and electrical faults — all detectable before failure.
Vibration signature analysis + MTBF regression surfaces failure windows 14–90 days before incident. Every algorithm is auditable.
average cost per hour of unplanned industrial downtime
Manufacturing plants hemorrhage capital on reactive maintenance. The data has existed for decades — the software was locked behind enterprise contracts.
OEE, MTBF, MTTR displayed per asset, per line, per facility. Drill from plant-wide to individual bearing in three clicks.
of maintenance organizations still use paper or spreadsheets
CMMS adoption lags because enterprise vendors price out everyone below Fortune 500 scale. That changes now.
40,000+ concurrent work orders across multi-site deployments. Priority routing, SLA tracking, spare parts forecasting — all open-source.
Three operators.
One platform.
Marcus Webb
Reliability Engineer, 800MW Gas Plant — Corpus Christi, TX
Tracking MTBF on aging turbines that haven't been replaced in 22 years
Bearing wear on Frame 7 turbines follows a degradation curve the OEM stopped publishing data for in 2008. Marcus was rebuilding failure models from scratch in Excel.
Maintain's vibration signature module ingests historian data from OSIsoft PI, applies a configurable degradation model, and surfaces failure probability windows per rotor stage.
14–90 days
Failure prediction window
< 3.2%
False-positive rate
61%
Avg downtime reduction
Priya Nair
Facility Director, 1,400-bed Academic Medical Center — Houston, TX
40,000 open work orders. One team. No margin for missed critical-equipment PMs.
Multi-site work order routing with regulatory compliance tracking. Automated PM escalation when SLA windows approach. Full audit trail exportable to Joint Commission format.
40,000+
Concurrent work orders
99.1%
PM compliance rate
−74%
Audit prep time
Darnell Okafor
Fleet Supervisor, Open-Pit Mine — Sudbury, Ontario
A mobile CMMS that works at 1,200m underground where LTE is a rumor
Haul truck PMs were being skipped because technicians couldn't access the CMMS 800 meters below surface. Paper forms meant 48-hour data lag.
Offline-first Progressive Web App caches the full work order queue, parts catalog, and equipment history locally. Syncs the moment connectivity returns. Works on Android rugged tablets.
< 8 sec
Offline sync delay
+43%
PM completion rate
48h → 0
Data lag eliminated
Every commit visible.
Every module forkable.
No black boxes. The algorithms that predict your turbine failures are the same ones you can read, audit, modify, and contribute to.
feat(failure-engine): add bearing wear regression model
fix(work-orders): resolve SLA escalation edge case at midnight UTC
feat(mobile): implement offline sync queue with conflict resolution
docs: add Frame 7 turbine vibration baseline reference
test(api): 100% coverage on asset hierarchy endpoints
Asset registry, hierarchy, attribute management
Predictive failure engine, MTBF regression, vibration analysis
Work order lifecycle, PM scheduling, SLA tracking
Offline-first PWA, sync engine, rugged device support
OEE dashboards, KPI reporting, export pipelines
3,654
Contributors
4,847
Commits
98.4%
Test coverage
One command.
Running in production.
No sales call. No 90-day implementation. No seven-figure license. Pull the container, configure your historian integration, deploy.
3,654
Contributors
14.2K
GitHub Stars
847M
Work orders processed
12,400
Active deployments
Zero friction. No email required.
No demo request. No sales qualification. No 30-day free trial that expires. Pull the container. Read the source. Deploy in production. That's the entire process.
The Rational Conclusion
The argument is complete.
Enterprise-grade failure prediction. Hospital-scale work order management. Mine-shaft offline capability. MIT license. Zero cost. One command.