Vision AI · Quality Inspection Software

SwitchOn DeepInspect® Zero Defects.
Full Speed.

Deploy AI-powered visual inspection on any production line in under 2 weeks — no data scientists, no coding, no line stoppage. Train new SKUs in 45 minutes. Trusted by global manufacturers to achieve zero-defect production.

99.95%+
Accuracy
1,000+
Parts / Min
45 Min
New SKU
150µm
Min Defect
Request Pricing ↓
DeepInspect® — Live Inspection Feed
AI bottle inspectionPASS
AI automotive inspectionFAIL
AI pharma inspectionPASS
AI electronics inspectionPASS
Auburn Hills, MI — US Integration Partner
2-Week Deployment
24/7/365 Support
Works With Your Existing Cameras & PLCs
99.95%+ Production AccuracyTrain New SKU in 45 Minutes1,000+ Parts Per Minute150 Micron Defect DetectionFDA 21 CFR Part 11No Coding Required8 Cameras Per Station50+ SKU ModelsSOC 2 Type IISiemens · Omron · Delta · Mitsubishi0% False Positives — POC ProvenAuto SKU SwitchingOn-Premise Edge AIISO Certified 99.95%+ Production AccuracyTrain New SKU in 45 Minutes1,000+ Parts Per Minute150 Micron Defect DetectionFDA 21 CFR Part 11No Coding Required8 Cameras Per Station50+ SKU ModelsSOC 2 Type IISiemens · Omron · Delta · Mitsubishi0% False Positives — POC ProvenAuto SKU SwitchingOn-Premise Edge AIISO Certified
01
Core Capabilities
Built for the Defects That Cost You the Most
99.95%+
Production Accuracy
Deep learning AI detects surface defects, label errors, particulate contamination, and assembly failures at accuracy levels no human inspection team can sustain across shifts. Proven across 1,000+ unique SKUs globally.
45 Min
New SKU Training — No Code
Capture fewer than 200 images of acceptable parts. The no-code interface guides operators step by step — no data scientists, no cloud dependency. A production-ready AI model in under an hour.
1,000+
Parts Per Minute
Up to 8 industrial cameras per station. 1.3 to 20 megapixels. Area scan, line scan, and thermal. Compatible with Basler, Baumer, Allied Vision, FLIR — whatever you already have on the line.
150µm
Minimum Defect Size
Detects scratches, dents, cracks, pits, discoloration, and inclusions down to 150 microns — on reflective, transparent, and complex-geometry surfaces where rule-based machine vision fails.
50+
SKU Models Per System
Deploy 50+ trained inspection models simultaneously. PLC changeover signal auto-loads the correct product model — zero manual intervention, zero line stoppage during transitions.
2 Wks
Kickoff to Live Inspection
Rev1 handles fixture design, camera mounting, PLC wiring, model training, and factory acceptance testing. One Michigan engineering team, start to production sign-off.
Rev1 Technologies — Auburn Hills, MI
We Don't Just Sell It. We Deploy It.
Rev1 is the US integration partner for this platform. We design the fixture, wire the PLC, train your team, and own the commissioning outcome — zero vision expertise required from your side.
(248) 707-2950
02
Deployment Process
Live on Your Line in 3 Steps / 2 Weeks
DeepInspect edge controller — industrial camera hardware setup
Day 1–2
Mount, Connect & Configure Hardware

Hardware-agnostic — works with any GigE Vision camera. Rev1 handles fixture design, camera mounting, and PLC wiring. TCP/IP and Modbus. Successfully integrated with Siemens, Delta, Omron, and Mitsubishi.

Any CameraBasler, Baumer, Allied Vision, FLIR — GigE / USB3 Vision
Any PLCSiemens, Omron, Delta, Mitsubishi — TCP/IP + Modbus
8 CamerasMaximum per station — 1.3 to 20MP, all shutter types
Day 2–3
Train the AI Model in Under an Hour

Fewer than 200 images of acceptable parts. No defect images required. The no-code training wizard guides your operators through every step. No data science team, no cloud connection. Production-ready in about 45 minutes.

<200Good-part images — no defects needed to start
45 MinTypical time from first capture to live model
No CodeOperator-friendly — no data science expertise needed
DeepInspect no-code AI model training interface
DeepInspect live production line AI inspection
Day 7–14
Go Live at Full Line Speed

100% automated inspection at up to 1,000+ PPM. Real-time pass/fail signals to your PLC for automatic rejection. Rev1 manages on-site commissioning, operator training, and post-deployment hypercare through production sign-off.

1,000+ PPMInspection speed — SKU, lighting, line conditions dependent
Real-TimePass/fail PLC signal — direct automatic rejection
10K ImagesOn-board archive for traceability and audit readiness
Auto SKU Switching
50+ Models / System
On-Premise — No Cloud
FDA 21 CFR Part 11
03
Industry Applications
Every Industry. Every Defect Type.
Automotive AI visual inspection — stamped parts, fasteners, welds
🏭 Automotive Manufacturing
Surface defects on stamped, painted, and cast components. Weld verification, fastener presence, brake pad, bearing, piston, and spark plug inspection. QA team reduced from 4 operators to 1 supervisor per shift at a Tier 1 supplier.
Pharmaceutical vial AI particulate inspection — FDA 21 CFR Part 11
💊 Pharmaceutical & Biopharma
Vial particulate contamination, capping defects, broken tablets, label compliance. 21 CFR Part 11 compliant — full audit trail. POC validated at 0% false positives and 0% false negatives across 6 particle classes.
Consumer goods packaging AI inspection — seal label fill level at 1000 PPM
🛒 Consumer Goods & FMCG
Seal integrity, cap placement, label verification, fill level, contamination at 1,000+ PPM. Defect rate reduced 0.3% → 0.05% at a global consumer goods manufacturer after replacing their legacy machine vision system.
Electronics PCB soldering AI inspection — SMT assembly verification
🔌 Electronics Manufacturing
PCB soldering defects, missing components, misaligned connectors, surface scratches on housings, and assembly completeness verification. Area scan and line scan cameras for board-level inspection at any throughput.
Precision metal bearing and fastener AI surface inspection
⚙️ Precision Metal Parts
Bearing, fastener, piston, and connecting rod surface inspection — regardless of part orientation. Handles reflective metal surfaces. Detects defects that make surface condition rejectable but pass dimensional gauging.
Glass bottle bore and fill level AI inspection — food beverage FMCG
🧪 Food, Beverage & Glass
Bore, wall thickness, chip, contamination, and fill-level inspection for bottles, vials, and jars on high-speed lines. Label placement and expiry date OCR verification — on-premise, no cloud required.
04
Proof of Concept Results
Real Performance. Validated On Your Line.
Biopharma sterile vial particulate inspection proof of concept
Biopharma — Sterile Vial Inspection
0% False Positives.
0% False Negatives.
Six Particle Classes.

A proof-of-concept for a global biopharma manufacturer — inspecting 20mL sterile vials for particulate contamination — achieved perfect validation across all six tested contaminant classes. Dual-camera optical assembly with diffuse backlight and ring light, integrated with a vacuum-gripper vial handler and PLC. POC ran at Rev1's Auburn Hills facility before on-site commissioning.

0%False Positive Rate
0%False Negative Rate
6Particle Classes Validated
2 WksPOC to Proposal
Consumer Goods — Soap & Personal Care
Defect Rate: 0.3% → 0.05%

A global consumer goods manufacturer replaced a prior machine vision system on a high-speed packaging line. Defect escape rate dropped 83% while maintaining 1,000+ PPM throughput — zero additional headcount.

83% Defect Reduction
1,000+ PPM Maintained
Zero New Headcount
Automotive — Tier 1 Stamping Supplier
4 Operators → 1 Supervisor Per Shift

Body-in-white surface inspection where traditional vision failed due to part placement variability. Pass/fail in under 1 second at line speed. False rejections down 40%. QA labor reduced 75%.

75% Labor Reduction
40% Fewer False Rejects
<1 Sec Pass/Fail
05
Analytics Platform
Monitor, Root Cause & Improve
Industry 4.0 Quality Intelligence
Every Inspection Becomes Actionable Data

Live rejection ratio trends. Defect Pareto analysis. Shift-over-shift comparisons. CPK/PPK reporting. Root-cause image archive. Multi-station, multi-facility aggregation — one dashboard, full factory visibility.

Multi-Station / Multi-Facility
10K Image Archive
REST API to ERP / MES
SOC 2 Type II Cloud
LIVE DeepInspect® Line 2 · HDPE · Shift A Sync 14:32:08 UTC LINE PPM (60 MIN) 1,186 GOOD (PASS) 14,892 NOT GOOD 71 THROUGHPUT / HOUR · GOOD vs NG ■ Good ■ NG 06 08 10 12 14 16 YIELD % · LAST 24 HOURS 98.2% 99.52% Defect Pareto (shift): surface scratch 42% · closure gap 28% · color 15%
Bottle inspection — DeepInspect visual quality on production line
🍶 P&G — Ariel HDPE Bottle, 40+ SKUs
8-camera DeepInspect system on an Ariel detergent line. Surface defects as small as 1cm detected at 50–130 PPM. 99.5% accuracy, 0.05% false positive rate. Real AI defect detection with red overlay shown above.
Machine vision piston inspection
AI based piston inspection system for automotive manufacturing
Train on fewer than 200 good images in under 45 minutes — fast shop-floor deployment. Up to 99.5% detection accuracy for automotive piston inspection.
Glass bottle bore inspection software
🍾 Glass Manufacturer — Bore Inspection, 300 PPM
40+ SKUs (50–350mm), trained with just 200 images. Detects defects over 500 microns — neck diameter variations, choked necks, and foreign particles. Eliminated 2 manual inspectors. 99.8% accuracy, 0 customer complaints.
06
Competitive Comparison
DeepInspect® vs. Leading Vision AI Platforms

When evaluating automated visual inspection, teams often look at traditional leaders like Cognex and Keyence, plus newer AI platforms. Here’s how DeepInspect® with Rev1 stands out — on capability, speed to production, and total cost.

Criteria DeepInspect®
via Rev1 Technologies
Cognex / Keyence Other AI Platforms
Deep learning (purpose-built)✓ NativeAdd-on or rule-basedVaries
Camera choice✓ Any GigE VisionVendor-lockedProprietary or bundled
Training data✓ <200 good images, no defectsHundreds–thousands or rule configOften defect data or labeling
Time to production✓ ~2 weeks with Rev112–16 weeks typicalSelf-serve only, no US integration
US integration & commissioning✓ Rev1 — fixture, PLC, sign-offIntegrator-dependentDirect only
FDA 21 CFR Part 11 (pharma)✓ Native, audit-readyCustom validationNot featured
Throughput1,000+ PPMLine-dependentLine-dependent
Relative investment●●●●●●●

● = most affordable   ●●● = mid   ●●●● = premium (no dollar amounts — contact Rev1 for your use case).

Why Teams Switch from Cognex & Keyence

Rule-based systems require machine vision engineers. When part geometry varies or lighting isn't perfect, false-positive rates spike and re-tuning is constant. Deep learning handles that variability without code changes.

How It Differs from Overview.ai & Elementary

Both are excellent platforms — but hardware-bundled and direct-to-manufacturer only. Neither provides a US engineering team to design your fixture, wire your PLC, and own the commissioning outcome. Rev1 does.

07
System Architecture
Edge AI + Cloud Analytics — On-Premise Inspection
DeepInspect system deployment architecture — cameras edge controller PLC cloud analytics
No Cloud Dependency for Inspection
All AI Runs On-Premise at the Machine

All inference runs locally on the DeepInspect edge controller. Inspection decisions happen in real time without a network connection — suitable for air-gapped and regulated environments. Cloud connectivity is optional, used only for analytics aggregation.

On-PremAll AI inference at the edge — no cloud required
8 CamerasArea scan, line scan, thermal per station
50+ SKUsAuto-switch on PLC trigger — zero changeover
REST APISAP, Oracle, MES, QMS integration
SOC 2 IICloud analytics — encrypted at rest and in transit
Siemens · Omron · Delta · Mitsubishi
10K Image Archive
Air-Gap Compatible
08
Technical Specifications
DeepInspect® Platform Specifications
Inspection Speed
Up to 1,000+ parts per minute
Detection Accuracy
99.95%+ production accuracy
Minimum Defect Size
150 microns (200µm recommended for 2D surface)
False Positive Rate
<0.5% — fully tunable per application
Camera Support
Area scan, line scan, thermal — up to 8 cameras / station
Camera Resolution
1.3 to 20 megapixels, global and rolling shutter
Supported Camera Brands
Basler, Baumer, Allied Vision, FLIR — all GigE / USB3 Vision
Training Images Required
<200 good-part images — no defect images to start
New SKU Training Time
~45 minutes — no coding, no cloud required
SKU Models Per System
50+ models simultaneously
PLC Integration
Siemens, Delta, Omron, Mitsubishi — TCP/IP and Modbus
ERP / MES Integration
REST API — SAP, Oracle, custom MES/QMS
Deployment Model
On-premise edge inference — no cloud required
Image Archive
10,000-image on-board save with inbuilt analytics
Licensing
Annual subscription — ~$10,000/station/year (up to 6 cameras)
Compliance
ISO · CE · FDA 21 CFR Part 11 · SOC 2 Type II
Deployment Timeline
~2 weeks kickoff to live inspection (Rev1 managed)
Support
24/7/365 — US-based integration via Rev1, Auburn Hills MI
09
Compliance & Certifications
Built for Regulated Manufacturing
Regulatory Certifications
Every Major Regulated Industry. One Platform.

Designed for pharmaceuticals, medical devices, food and beverage, and aerospace — where inspection traceability, data integrity, and audit readiness are non-negotiable.

CE Certified ISO Certified FDA 21 CFR Part 11 SOC 2 Type II
CE Certified
ISO Certified
FDA 21 CFR Part 11
SOC 2 Type II
21 CFR Part 11 — Electronic Records

Full audit trail, electronic signature support, and access control logging for FDA-regulated pharmaceutical and medical device manufacturing. SSO included in all subscription tiers.

SOC 2 Type II — Cloud Security

DeepInspect Analyze cloud platform operates on SOC 2 Type II infrastructure. All inspection data at rest and in transit is encrypted to enterprise security standards.

Air-Gapped Deployment

All AI inference runs locally — the platform operates without network connectivity, suitable for classified and regulated environments.

10
Frequently Asked Questions
Common Questions About Vision AI Inspection
Do I need to replace my existing cameras?
Not necessarily. The platform is hardware-agnostic and works with most industrial cameras using GigE Vision or USB3 Vision standards — Basler, Baumer, Allied Vision, FLIR, and others. Rev1 assesses your existing hardware during scoping and recommends the most cost-effective path. Where new hardware is required, we spec and source it as part of the total solution.
How long from contract to live inspection?
Typically 2 weeks from kickoff to live production inspection — including hardware mounting, camera setup, model training, PLC integration, and factory acceptance testing. Complex multi-station deployments may take 4–6 weeks. Rev1 manages the full project scope from engineering through commissioning and sign-off.
Can my operators train the system themselves?
Yes. The no-code training interface is designed for production floor teams — not engineers. Capture fewer than 200 images of acceptable parts, run the training wizard, and the model is production-ready in about 45 minutes. No Python, no cloud, no machine learning expertise required. Rev1 provides full operator training during commissioning.
What types of defects can it detect?
Surface defects (scratches, dents, cracks, pits, discoloration, porosity), label and OCR verification (expiry dates, barcodes, lot numbers), thermal defects (seal integrity, cap presence), object counting and assembly completeness, and particulate contamination in liquids and vials. The AI learns your specific defect profile — not a generic definition of "defect."
How does it handle product changeovers?
Up to 50+ trained inspection models deploy simultaneously. When your PLC sends a changeover signal, the system automatically loads the correct product model — 100% inspection coverage maintained through the transition with zero manual intervention and no line stoppage.
Is it compliant for pharmaceutical manufacturing?
Yes. FDA 21 CFR Part 11 compliant with full electronic audit trail, user access controls, and SSO. CE and ISO certified. A recent POC achieved 0% false positives and 0% false negatives across six tested particle classes for a biopharma manufacturer. Contact Rev1 to discuss your IQ/OQ/PQ requirements.
What does pricing look like?
One-time hardware and integration investment (scoped per application) plus an annual software subscription starting around $10,000 per station per year. A station license supports up to 6 cameras. Subscription includes platform access, model hosting, software updates, and 24/7/365 support. Rev1 provides a detailed itemized quote after an initial scoping call.
What software pairs well alongside vision AI?
For dimensional metrology, PolyWorks Inspector pairs naturally with automated visual inspection for full-coverage QC. For reverse engineering, Geomagic Design X is available through Rev1. See our full software catalog.