field capture, fed straight into the platform

Smart glasses → structured maintenance data.

R&D Concept · Smart Glasses

A technician wears NixVision glasses on a walk-down, with NixAssist running on their phone. What they see and say is recorded hands-free and sent to Neuralix DLT, where it's tagged, checked against live SCADA, scored by health models, and logged automatically. Alerts and guidance come back to the technician through audio and the NixAssist app. Each inspection ends as structured maintenance data and a labelled dataset.

EDGE DATA CAPTURE 01 NixVision capture ↔ live feedback 02 NixAssist mobile · voice, chat, photos Neuralix DLT digital twin · ingestion, AI models & system of record — one platform INGESTION & AI PIPELINE Ingest frames · metadata Vision Tagging components · defects SCADA Correlate vs live measurements Health ML Models history → anomaly DIGITAL TWIN · SYSTEM OF RECORD Time-series+ live SCADA Maintenance& lifecycle log Image / videoarchive Labelleddataset lake SINGLE SOURCE OF TRUTH AUTO-LOG SCADA + HISTORY
capture source data flow

How the flow works

Two domains: capture at the edge, and Neuralix DLT — where ingestion, the AI models and the system of record all live. Hands-free for the technician; everything past the glasses runs inside the platform.

1

Capture at the edge

NixVision glasses record photos and video of equipment on a normal walk-down, while NixAssist on the phone adds voice notes, forms and chat — no clipboard, no interrupting the work.

NixVision + NixAssist
2

Into the data pipeline

Footage syncs and is split into frames with metadata attached: timestamp, location, and the asset it belongs to.

ingest · frames · metadata
3

Tag & analyze images

Vision models detect equipment, components and visible defects — corrosion, leaks, gauge readings, loose parts — and label each frame.

vision · detection · labels
4

Correlate with SCADA

Tagged visuals are aligned against live SCADA measurements pulled from Neuralix DLT, so what's seen is matched to what's measured.

DLT · live measurements
5

Health ML in the background

Models weigh the visuals and sensor data against historical records, past images and logs to score system health and surface anomalies.

anomaly · diagnostics
6

Auto-log back to DLT

A structured maintenance record — media attached — is written straight back into the platform, and NixAssist agents notify or escalate to the right crew. No manual data entry.

auto-log · notify · zero entry

NixAssist, fully integrated

The glasses don't ship as a standalone gadget — they plug into NixAssist, the conversational layer already wired into DLT. NixAssist is the connective tissue across the whole loop.

Co-capture on the phone

NixAssist's mobile app pairs with the glasses, adding voice notes, chat and forms alongside the hands-free POV footage — one capture session, two devices.

Agents orchestrate

The same Azure-OpenAI agents that power NixAssist drive the pipeline: image-to-workflow, keyword triggers, and multi-step follow-ups across every stage.

Already speaks to DLT

NixAssist is multi-tenant and connected to the customer DB today, so write-back, notifications and escalations land in the platform with no new plumbing.

Why it matters

The capture loop pays off twice: better maintenance today, and a compounding proprietary dataset for tomorrow.

Maintenance logging

Every inspection becomes a complete, searchable record — automatically, with evidence attached.

Dataset collection

A growing proprietary library of field imagery tied to real sensor data — a moat that compounds over time.

Auto-labelling

SCADA correlation turns raw footage into labelled training data without a manual annotation team.

Video knowledge base

Captured walk-downs become a reusable reference for training, audits and root-cause review.

“The work a technician already does becomes the data that trains the next model — and the maintenance record writes itself.”