← Platform Overview

Edge Intelligence

TwinEdge OS

A full industrial operating system running on your hardware. Multi-protocol connectivity, real-time ML inference, and local dashboards -- all with zero cloud dependency.

PROTOCOLSOPC UA:4840Modbus:502MQTT:1883Sparkplug BMQTTS7comm:102EtherNet/IP:44818TWINEDGE OSOPC UA HubProtocol Unifier:4840OrchestratorSubscription Engine:5000ML EngineONNX RuntimeStorageLocal DB + ParquetAlertsMulti-ChannelDashboardHMI + WebSocket:3000OUTPUTSDashboard HMI:3000Cloud SyncMQTTAlert ChannelsSMS/EmailData ExportParquettwinedge@hub-pro:~$ twinedge status --all ● opcua-server active (:4840) 10 protocols ● orchestrator active (:5000) subscription-based ● ml-inference active (:8000) 3 models loaded ● storage-service active (:5001) 2.4 GB / 16 GB ● alert-service active (:5003) 4 channels ● dashboard active (:3000) PWA + offline

Protocol Support

10 industrial protocols — 7 production-ready, 3 in beta. See full gateway specs →

ProtocolIntegrationLatency
OPC UANative<10ms
Modbus TCPNative<5ms
Modbus RTUNative<8ms
MQTT 3.1.1/5.0Native<2ms
Sparkplug BNative<2ms
EtherNet/IP (CIP)Native<10ms
S7comm / S7comm+Adapter<15ms
BACnet/IPBeta<10ms
PROFINETBeta<15ms
DNP3Beta<20ms

Edge Capabilities

On-Device ML Inference

ONNX Runtime executes anomaly detection, RUL prediction, and efficiency models directly on the edge device. No cloud dependency.

Local Dashboards

Real-time dashboards served from the edge. Access via any browser on the local network -- no internet required.

100% Offline Operation

TwinEdge OS runs indefinitely without connectivity. Data buffers locally and syncs when the connection returns.

Store-and-Forward

Embedded time-series storage with automatic Parquet export. Buffered telemetry syncs to cloud on reconnect without data loss.

Deploy Edge Intelligence Today

Ship a TwinEdge Hub to your site, plug it in, and start monitoring in under an hour.