01Home 02Platform 03Hardware 04Industries 05Request a pilot

One OS. Every asset. Any vendor.

KinetiqOS is not another dashboard. It is the layer that runs the plant; ingesting every signal, modeling every asset, predicting every failure mode it has been trained on, and putting the next action in front of the right human.

ArchitectureOn-prem first · Docker
Data planeMQTT → Kafka → Analytics
IntelligenceDSP + Autoencoder AI
Vendor lock-inNone
01 The signal pipeline

From the bearing
to the boardroom.

One unbroken chain carries every sample from the machine surface to a decision. No exports, no swivel-chair integration; the pipeline is the product.

SENSORS PLC · DCS · LACC001 OPC-UA · MODBUS · 4–20MA EDGE GATEWAY PROTOCOL BRIDGE TRANSLATE → MQTT MQTT BROKER · TLS TOPIC CONTRACTS KAFKA STREAM BACKBONE REPLAYABLE · DURABLE ANALYTICS DSP FEATURES FFT · ENVELOPE · KURTOSIS AI AUTOENCODER ANOMALY INFERENCE APPS ACT & OPERATE DASH · ALERTS · FORMS STAGE 01STAGE 02STAGE 03 STAGE 04STAGE 05STAGE 06STAGE 07 OUTCOMES FEED BACK · EVERY INTERVENTION RETRAINS THE MODEL OF THE PLANT
02 The platform, layer by layer

An OS is a stack.
Here is ours.

Like any operating system, KinetiqOS abstracts the hardware below it so applications above it can be simple, portable, and fast to build.

L-1
LAYER 01 · ACQUISITION

Sense

Kinetiq's LACC001 nodes capture triaxial vibration and temperature at up to 25.6 kHz. Where instrumentation already exists, edge gateways bridge OPC-UA, Modbus, 4–20 mA, IO-Link, BACnet, and EtherNet/IP into the same MQTT stream; one ingestion contract for the whole plant.

LACC001 native node7+ industrial protocols2 Hz – 25.6 kHz ODR
L-2
LAYER 02 · CONTEXT

Understand

Signals are meaningless without context. KinetiqOS normalizes every stream and binds it to an asset in a plant–area–line–asset hierarchy, so a 4 mm/s reading on a kiln girth gear is judged as a kiln girth gear; not as an anonymous number.

Asset hierarchySignal normalizationFleet-wide context
L-3
LAYER 03 · INTELLIGENCE

Predict

Server-side analytics compute RMS, FFT, envelope spectra via Hilbert transform, crest factor, kurtosis, and skew on every node. On top, an autoencoder learns each asset's healthy fingerprint and flags reconstruction anomalies; catching faults with no labeled failure data required.

FFT spectraEnvelope analysisAutoencoder anomaly detection
L-4
LAYER 04 · ACTION

Act

A prediction nobody acts on is trivia. KinetiqOS routes anomalies into alerts, diagnostic cases, live dashboards, and configurable forms and workflows; with exports and APIs that push the verdict into the maintenance systems you already run.

Live dashboardsAlertingForms & workflowsExports & APIs
L-5
LAYER 05 · ECOSYSTEM

Extend

The OS is a floor, not a ceiling. Connectors bring new sensor vendors into the data model; apps add capabilities; energy, quality, production; on top of the same plant model, without re-platforming.

Connector SDKApp modelPartner-ready
03 Analytics & AI

Watch the autoencoder's job, live.

The healthy baseline is learned, not configured. Inject a fault below; the spectrum shifts, crest factor climbs, and the state flag trips. That flag is what your maintenance planner sees, weeks early.

KINETIQ ANALYTICS · INFERENCE VIEW Model running NOMINAL
Time domain; acceleration, g
Frequency domain · FFT magnitude
Velocity RMS
-
Crest factor
-
Kurtosis
-
Temp
-
State
-
Fault physics, visualized: imbalance raises the 1× shaft harmonic; an outer-race defect rings a high-frequency resonance in an impulse train; envelope analysis digs it out of the noise.
04 The plant model

Every asset,
one tree.

Risk rolls up from the bearing to the boardroom. A hot asset colors its line, its area, and its plant; so attention always lands where the physics says it should.

PLANTWorks 02 · Cement
Areas
RAW MILL12 assets
KILN LINE18 assets
FINISHING9 assets
Assets; kiln line
KILN DRIVE1× rising · 4.1 mm/s
GIRTH GEARMesh wear · crest 6.2
ID FAN1.8 mm/s
COOLER FAN2.1 mm/s
Fleet risk rolls upward in real time
05 Running KinetiqOS
DEPLOY / 01

On-prem first

The full platform ships as Docker services and runs inside your plant network. Your process data never has to leave the site.

DEPLOY / 02

Edge inference

Bandwidth-limited or remote sites can run detection on the gateway itself and ship only verdicts upstream.

DEPLOY / 03

Secured by design

Secured REST and WebSocket APIs, TLS transport on the broker, and role-based access from operator to administrator.