Deployment Process
From install to live inspection in under a week
No custom infrastructure. No month-long integration. Gaugegrove ships pre-configured edge hardware and a browser-based setup wizard that walks your team through every step.
Week 1, Day 1-2
Mount and connect cameras
Industrial cameras mount at your inspection station on adjustable brackets included with the kit. Connect to the Gaugegrove edge unit via standard Ethernet. The setup wizard detects each camera automatically and walks you through focus and exposure calibration.
Compatible with any GigE Vision or USB3 Vision industrial camera.
Week 1, Day 2-4
Train on your specific defect classes
Upload or capture examples of each defect class specific to your parts. The platform fine-tunes a pre-trained vision backbone on your data. For common defect classes (scratch, crack, edge roll, missing component), starter models are available and require as few as 50 examples per class.
Most customers reach target accuracy within 4-6 hours of training data collection.
Week 1, Day 4-5
Validate in shadow mode before going live
Run the model in shadow mode: it inspects every part and records results alongside your human inspection team, but triggers no automated reject signals. Compare AI detections against known defective parts from your validation set and tune confidence thresholds until precision and recall hit your targets.
Typical shadow mode validation: 1-3 shifts with a known-defect validation set.
Week 1, Day 5+
Go live: automated inspection at line speed
Flip from shadow mode to live. Reject signals route to your PLC or mechanical diverter. Defect events push to your MES via webhook automatically. The dashboard shows every pass/fail in real time. Your team monitors trends, not individual parts.
Average go-live timeline: 4-7 days from first camera mount to first live inspection.
See the setup wizard in a live walkthrough
We'll demo the full deployment flow on a sample inspection station.