Step 01Week 1–2
Connect
We read from your existing PLCs, SCADA, and machine sensors via OPC-UA or MQTT. Edge-only deployment available for air-gapped sites.
Industry & Operations
factoryNET
Catch line failures before they catch you. AI-powered monitoring of factory floor operations and production quality control.
BuiltforDiscrete & process manufacturers running OPC-UA / MQTT machinery.
factorynet.dunavnet.eu — Automat novog emajla
How it works
From sensor stream to actionable alert in four steps.
factoryNET sits beside your existing PLC, SCADA, and MES infrastructure — it does not replace it. Pilot one line in 4–6 weeks. Plant-wide in 3–6 months.
Step 02Week 2–4
Baseline
Models learn your line's normal behaviour from historian data and live sensor streams. No manual tagging required for most signatures.
Step 03Week 4–6
Detect
Anomaly detection surfaces small deviations that precede the expensive ones. Alerts to dashboard, email, Teams, or your MES via REST API.
Step 04Ongoing
Improve
Operator feedback closes the loop. Models continuously adapt. You see month-over-month reductions in unplanned stops and scrap.
From stream to alert — the journey, end to end
Solution portfolio
See exactly what your team sees on the floor.
From live production dashboards to field readings on the line and the anomalies factoryNET catches before they become scrap — these are the screens your operators, supervisors, and plant managers work with every shift.
factorynet.dunavnet.eu — Cookware detection · Litvanska peć
Integrates with your stack
Built to fit. Not to rip and replace.
We read from what you have, output structured anomaly events, and speak the protocols your MES, ERP, and BI tools already understand.
OPC-UA
MQTT
Modbus TCP
InfluxDB
Don't see your system? We've integrated with 30+ industrial systems across our deployments. Bring it up during the discovery call.
How factoryNET fits
Reference architecture
Your environment
Plant floor
PLC · SCADA · sensors
OPC-UA · MQTT
factoryNET
Edge gateway
factoryNET runtime — models run locally
Anomaly events only
factoryNET
factoryNET cloud
Dashboard · alerts · APIs
REST · webhooks
Your environment
Your existing systems
MES · ERP · BI · Teams
From the field
Where the sensors actually live.
Real installs, real plants. Sensors on the line, the edge gateway in the cabinet, the team on site.
Deploy three ways
Pick the model that matches your security and ops posture.
SaaS
Hosted by us · EU region
Edge gateway on-site, dashboard in our EU cloud. Fastest start. Monthly billing. ISO 27001-controlled.
- Live in 4–6 weeks
- Monthly · per line
- EU data residency
Managed AzureMost chosen
Your tenant · we operate
Deployed inside your Azure tenant. Your data residency, your compliance posture, our 24/7 operations team.
- Your tenant
- We run the ops
- Co-managed SLA
Custom / on-prem
Air-gapped possible
Fully on-premise. Custom integrations. Optional via the DunavNET Innovation Studio engagement model.
- Air-gapped option
- Custom integrations
- Studio engagement
Customer story
What 90 days looked like on one European auto-parts line.
European automotive-parts manufacturer · deployed 2025
"In the first 90 days, factoryNET caught two bearing failures before they took the line down. Each one would have cost us a shift of production."
Plant operations director
380-line discrete manufacturer · 2 sites · Czechia + Slovakia
38%Reduction in unplanned downtime, first 90 days
4Bearing failures predicted vs. found
22%Reduction in scrap from quality drift
8 moPayback period at typical utilisation
Common questions
What buyers usually ask before they sign.
Not usually. We work from your existing PLC, SCADA, and historian data via OPC-UA or MQTT. If a specific signal is missing for your use case, we'll flag it during the discovery call and scope the additional instrumentation separately.
Only anonymised anomaly events — not raw sensor data. In the SaaS model, edge inference runs on-site; only structured alerts are sent to the cloud. In the on-prem model, nothing leaves your network at all.
Models are trained on your own historian data inside your deployment boundary. We do not pool data across customers. The initial baseline builds in 2–4 weeks; the model adapts continuously from operator feedback.
Pricing is per monitored production line per month. Most pilots cover 1–3 lines. Full plant-wide deployments are quoted on scope. We publish indicative ranges during the discovery call once we understand your line count and protocol mix.
Generic platforms give you infrastructure; factoryNET gives you outcomes. We pre-package the anomaly models, the protocol adapters (OPC-UA, MQTT, Modbus), and the operator UX — so your team is not building pipelines, they are reading alerts. DunavNET also handles the ongoing model tuning, not just the software licence.
Related solutions
