Connectivity, an asset model, a pipeline designer, operator screens, alarming, a historian and fleet management — in one install, sharing one namespace, licensed once.
Speak every protocol on the plant floor without a middleman.
OPC UA, MQTT with Sparkplug B, Modbus, SNMP, raw TCP/UDP and REST — through one pooled, buffered gateway.
A conformant primary host, not MQTT with a schema bolted on — birth and death certificates, sequence tracking and rebirth.
Poll and trap-receive from switches, UPSs, gateways and PDUs — and compile a vendor MIB into usable tags instead of hand-typing OIDs.
REST for configuration and data, WebSocket for push. The product's own UI is built on them, so they cannot quietly rot.
Turn raw tags into an asset model your whole business can read.
One live tree of everything you measure, with typed asset templates instead of forty copies of the same pump.
Compose pipelines, screens and logic without leaving the browser.
Not a chatbot bolted onto the docs. It reads your live namespace and calls real platform tools — bounded by the same permissions as the person asking.
Drag-and-drop dataflow on a real DAG engine — reads, writes, transforms, joins, database queries, scripts and schedules.
57 built-in widgets — charts, tables, forms, maps, P&ID and HMI symbol libraries, vector drawing, media and documents.
Load a glTF model of the line and bind its parts to live tags — so the twin is driven by the plant instead of drifting away from it.
Python for the logic no canvas will express, plus a sandboxed JS expression engine — both isolated from the runtime.
Run it in production: alarms, history, security, fleet management.
Deadband-filtered storage with rollups, retention policies, drift correction and a store-and-forward sink.
Alarms defined on any namespace node, evaluated by a dedicated service, with an audited acknowledgement trail.
The link will drop. Samples buffer locally and reconcile on reconnect, with original timestamps intact.
Granular role-based permissions and an audit log that records configuration changes and operator writes.
No connector marketplace and no per-driver licence. Everything below is in the binary you already downloaded. Siemens S7 and EtherNet/IP are not supported yet — they are on the roadmap.
Client with subscription batching and pooled sessions.
MQTT client, pooled and buffered, with automatic reconnect.
Primary host: birth/death certificates, sequence and state.
Master with scan grouping and explicit byte-order control.
v1/v2c/v3 polling and traps, with a built-in MIB compiler.
Raw sockets for the devices that speak nothing standard.
Inbound routes and outbound calls, both scope-controlled.
Push curated data where the rest of the business already works — with schema, units and quality codes intact.
Read, write and historise
Read, write and historise
Existing MES and ERP schemas
Embedded, zero-admin deployments
Document sink for batch and event data
External time-series destination
Six patterns that cover most of what teams build in their first year.
QUBIQ watches machine state directly, times every stop, and prompts the operator to classify anything above a threshold. Availability, performance and quality roll up by line, shift and product automatically.
Pull from meters over Modbus and SNMP, normalise units in a pipeline, and divide by good-units-produced from the line. Cost per unit becomes a number operators can actually influence during the shift.
High-rate condition data is captured at the edge, aggregated, and compared to per-asset-type baselines. When a bearing drifts, the alarm carries the asset, the work order template and the last thirty days of trend.
Every batch gets a record built from the namespace: setpoints, actuals, deviations, operator signatures and out-of-spec events, with a signed audit trail suitable for a regulated review.
Because assets are typed in the namespace and the 3D viewer binds to that same namespace, the twin and production cannot drift apart. Visibility, transforms, materials and animations all follow live values.
Network and power infrastructure is where plant outages usually begin, and it is almost never in the SCADA system. SNMP polling plus the MIB compiler puts it there without transcribing OIDs by hand.
Every tour, deep dive and recorded session — filtered by whichever one you came for.
Bring an OPC UA endpoint or an MQTT broker to a 30-minute call and we will connect it live. If it does not work, that is useful information too.