Request a briefing
■ Products

Starting with vision. Built for the boundary.

Vision Analytics is our first product. The wider platform follows, under one rule: your data stays inside your boundary.

Vision Analytics is available for pilots now. OT Detection and Enclave AI are in development.

01 Vision Analytics · first product

AI visual inspection and PPE detection at the edge.

Connect existing IP cameras to an Ironmear edge node. Models run on site to detect surface defects, PPE gaps and line stoppages, and send alerts to operators and your MES or SCADA. No video leaves the plant.

Illustrative: an edge camera inspects parts on a conveyor; one part is flagged as a defect.CAM-01DEFECT · SCRATCH 0.97ILLUSTRATIVE · NOT REAL DATAEDGE NODE · ON-PREM
Fig. 1 — Edge inspection, illustrative
CapabilityDetailStatus
Defect detectionSurface defects, missing features, assembly errorsAvailable for pilots
PPE & safetyHelmet, vest and zone compliance alertsAvailable for pilots
Line monitoringStoppages, jams and throughput anomaliesAvailable for pilots
Edge deploymentOn-prem GPU node, cameras over RTSP/ONVIF, no cloud egressAvailable for pilots
■ Platform roadmap

Beyond vision: one security platform for industrial and defense.

The same on-prem foundation extends to control-network threat detection and air-gapped AI for defense teams. Both are in development. We will share timelines when they are firm.

02 OT Detection

See every command on the control network.

Deployed on a hardened appliance or VM at the industrial DMZ or cell level. It parses industrial protocols, builds a baseline of normal behaviour per asset, and alerts on deviations: unexpected writes, new devices, firmware changes, out-of-range process values.

Illustrative chart: observed process signal against learned baseline band, with one flagged deviation.DEVIATION · MODBUS FC16write outside baselineILLUSTRATIVE DATA · NOT A REAL INCIDENTt →
Fig. 1 — Illustrative
CapabilityDetailStatus
Asset inventoryPassive discovery of PLCs, HMIs, engineering workstationsIn development
Protocol analysisModbus/TCP, DNP3, S7comm; OPC UAIn development
Behavioural baselinePer-asset models of commands, timing, process valuesIn development
IntegrationsSyslog, CEF, common SIEMsIn development
03 Enclave AI

Capable models where the network ends.

A self-contained AI stack for disconnected and classified environments: document search and question-answering over your holdings, summarisation, and analyst tooling, all running on customer hardware with no external dependencies.

Enclave: model weights, retrieval index and documents in one sealed environment, updates delivered by signed offline media.CLASSIFIED ENCLAVE · AIR-GAPPEDMODEL WEIGHTSpinned, hashedRETRIEVAL INDEXyour documentsINFERENCE · ACCESS CONTROL · AUDITANALYST UIAPI (local)SIGNEDUPDATE MEDIAoffline
Fig. 2 — Enclave components
CapabilityDetailStatus
Offline inferenceOpen-weight models served locally on GPU serversIn development
RetrievalSearch and Q&A over local document stores with source citationsIn development
Access controlRole-based access, per-collection permissions, immutable audit logIn development
Update pathSigned, hash-verified offline update bundlesIn development