Protex AI
Edge-computing workplace-safety computer vision, doing behavior-safety analysis without footage leaving the plant
Protex AI is a workplace-safety-tech company headquartered in Dublin with a US base in Boston. Its biggest difference from similar products lies in architecture: all image analysis is done on edge devices inside the plant, raw footage isn't uploaded to the cloud, and only the analyzed event data leaves the site. For European customers with strict privacy regulations, this isn't a bonus but a necessity.
Features and application scenarios
The product turns existing CCTV into a source of "on-site intelligence," and the detection scope isn't just workplace safety — PPE wearing, ergonomic-posture risk, and vehicle control are the safety side; the other half is the operations side, like production-line bottlenecks, path congestion, and process deviation. This dual-track design is smart, because workplace-safety budgets are usually hard to get, but as long as it can also give operational-efficiency numbers, the procurement case becomes much easier. The system can connect with existing EHS management systems and BI tools to automatically produce reports and alerts. They announced completing a $36M Series B in January 2025, with a client list including DHL, Amazon, Sysco, Coca-Cola, and General Motors.
Suited to logistics, manufacturing, retail, food and beverage, and warehousing, especially units within multinational enterprises needing to meet GDPR or similar personal-data norms. If Taiwan's tech plants and logistics centers have a European parent company or client audit requirement, Protex AI's edge architecture is a clear advantage — the effort saved when explaining "footage doesn't leave the plant" to legal is worth more than the technical spec itself.
Main features
- Edge-computing architecture, footage processed on site, raw video not uploaded to the cloud
- Connects to existing CCTV, no need to replace camera hardware
- Safety detection: PPE compliance, ergonomic-posture risk, human-vehicle path control
- Operations detection: production-line throughput bottlenecks, path congestion, process deviation
- Integration with existing EHS management systems and BI tools
- Automated reports and real-time alerts
Common uses
- Forklift-and-pedestrian conflict early warning in logistics warehousing
- Automatic auditing of production-line PPE compliance
- Warehouse path-congestion and throughput-bottleneck analysis
- Cross-plant workplace-safety-event statistics and trend reports
- Image-analysis scenarios needing to meet privacy norms like GDPR
Key Features
- Edge-computing architecture, footage processed on site, raw video not uploaded to the cloud
- Connects to existing CCTV, no need to replace camera hardware
- Safety detection: PPE compliance, ergonomic-posture risk, human-vehicle path control
- Operations detection: production-line throughput bottlenecks, path congestion, process deviation
- Integration with existing EHS management systems and BI tools
- Automated reports and real-time alerts
Pros
- The edge-processing privacy architecture is a real differentiator, not marketing packaging
- Gives both safety and operations value, making internal proposals easier to sell to finance
- Client list has big-enterprise backing like DHL, Amazon, and GM
- Reuses existing cameras, so hardware investment is relatively controllable
Cons
- Requires deploying edge-computing devices inside the plant, so server-room and network configuration must be inventoried first
- Pricing not public
- Highly overlapping features with Intenseye and Voxel, easily bogging you down in spec comparison when choosing
- The labor-relations issue of employee-behavior monitoring equally exists and won't disappear just because footage doesn't go to the cloud
Use Cases
- Forklift-and-pedestrian conflict early warning in logistics warehousing
- Automatic auditing of production-line PPE compliance
- Warehouse path-congestion and throughput-bottleneck analysis
- Cross-plant workplace-safety-event statistics and trend reports
- Image-analysis scenarios needing to meet privacy norms like GDPR
Editor's Note
Editor's note: I think Protex AI's smartest part isn't the technology but the product positioning. It's always hardest for a safety department to get budget — because its performance is 'things that didn't happen,' which is hard to quantify. Tying operational-efficiency analysis into the same system gives the client's safety manager a reason to talk to the CFO. As for the edge architecture, in Europe it's the ticket to entry, not a bonus; but if you're in Taiwan and happen to need to satisfy a European client's audit, this ticket saves you a lot of trouble.
FAQ
What does 'edge computing' actually mean?
It means image analysis runs to completion on a machine inside your plant, and only the result (e.g., '10:32, someone at Gate 3 without a hard hat') is sent out, with the raw footage staying on site. In contrast, a cloud architecture uploads the image stream to an external server for analysis. The difference is data residency and bandwidth cost.
How do I choose between it and Intenseye?
Protex AI's edge-privacy architecture is more advantageous for units under GDPR or strict personal-data audits; Intenseye puts more into ergonomic-posture analysis and its own industrial hardware. If privacy compliance is your top priority, Protex AI is worth prioritizing.
If footage doesn't go to the cloud, how do I see reports?
The analyzed event metadata (time, location, event type, severity) still syncs to the management platform, and you can see statistics and trends on the dashboard — there just won't be raw footage in the cloud. When you need to pull up the picture, there's usually a controlled access mechanism.