Factories & smart manufacturing

Factories.
AI at the line,
not in the cloud.

Eldric runs at the line — workstation or single server next to the PLC. Predictive maintenance, alarm triage, OEE analytics, shift-handover notes. Soft real-time tag-to-action — about 50 ms on the demo cluster. The strongest production-tested workload Eldric ships today.

Talk to us All use cases
Workloads

What people in this sector typically run.

Predictive maintenance

Live tag values flow into matrix memory; an inference worker runs anomaly detection; maintenance scores emit on the line.

Industrial protocols

OPC-UA, Modbus TCP/RTU, MQTT Sparkplug B in the IoT worker. Store-and-forward through disconnect windows.

xLSTM forecasting

Time-series forecasting daemon (port 8884) for trajectory, wear, and yield prediction on sensor streams. Linear in sequence length.

Shift-handover summarisation

End-of-shift digest of alarms, deviations, set-point changes. Drafted for the incoming shift supervisor.

OEE analytics

Availability, performance, quality roll-up. Compared against plan. Aligned to the historian's time-series.

Edge deployment

Pi 4 to rack-mounted server — same binary. Modules per node. Air-gap installs for clean networks.

Compliance & isolation

What we explicitly support.

ISO 27001 documentation-friendly IEC 62443 (industrial network security) Audit ledger hash-chained Air-gap ready

Honest scope.

Soft real-time (10–100 ms) and near-real-time, yes. Hard real-time, deterministic control, RTOS-grade — no, that stays the PLC's job. Eldric sits above the PLC layer.

Next step.

Write to office@eldric.ai. Tell us what you are trying to do; we will tell you whether Eldric is a fit and, if not, what would be.