The Platform

Camera-to-MES in a single integrated system

Gaugegrove combines industrial camera orchestration, on-premises AI inference, and structured defect reporting into one system that runs on your floor without cloud dependency.

16
Cameras per edge unit
<3ms
Per-frame latency
99.8%
Accuracy

Architecture

Four stages, zero cloud dependencies

Stage 01
Capture

Industrial cameras (GigE Vision, USB3) trigger on encoder pulse. Up to 16 cameras per edge unit, synchronized to sub-millisecond precision.

Stage 02
Infer

GPU-accelerated model runs fully on-premises. Processes each frame in under 3ms. No frame leaves the facility, no cloud connection required.

Stage 03
Decide

Pass/fail threshold evaluated against per-class confidence scores. Configurable per defect type. Reject signals output to PLC or mechanical diverter.

Stage 04
Report

Structured defect event pushed to your MES via webhook. Dashboard updated in real time. Shift reports generated automatically.

99.8%
Inspection Accuracy
Across active deployments
<3ms
Frame Latency
On edge GPU hardware
16
Cameras per Unit
Sub-ms sync
72%
Escape Reduction
Reported by early adopters
Gaugegrove camera mounted at factory inspection station

Technical Specs

Hardware and software specifications

Camera Support
GigE Vision, USB3 Vision
Up to 16 cameras per edge unit, major industrial sensor brands supported
Inference Latency
<3ms per frame
Measured on Gaugegrove G2 edge unit at 5MP resolution
Edge Hardware
Included with subscription
Industrial-grade GPU unit, DIN rail mount, IP54 rated, 0-55C operating range
MES Integration
Webhook / REST API
SAP, Plex, Infor CloudSuite, and any REST-capable target
Data Security
VLAN + TLS at edge
Camera images processed on-prem. Configurable data residency and retention policies.
Dashboard
Browser-based, no install
Real-time pass/fail stream, defect heatmaps, shift summaries, confidence trends

See the platform running on parts like yours

We'll build a live demo using your part geometry and defect classes.