Global – October 29, 2026 -- The deviceWISE Intelligence Suite now integrates NVIDIA Metropolis Video Search and Summarization (VSS) Blueprint, feeding camera detections directly into live machine data and automated response workflows on production lines.
Platform connects PLCs, robots and CNCs to video analysis
The suite deploys industrial AI agents at the operational edge and links PLCs, robots, CNCs and enterprise systems into one real-time operational picture. NVIDIA VSS Blueprint adds accelerated computer vision, vision-language models including NVIDIA Cosmos, and large language models to analyze live and recorded video.
The Model Context Protocol (MCP) and NVIDIA NIM microservices connect the visual insights to machine state and process history. NIM matches each task to the most effective model to speed agent deployment.
Prebuilt VSS skills let developers build video agents from natural language
VSS skills are a catalog of reusable building blocks covering core video analytics functions. Developers describe a need in natural language, and a coding agent such as Codex or Claude assembles the skills into a working video analytics agent. The NVIDIA NemoClaw blueprint provides a reference for deploying these agents with guardrails in production.
Quality inspection and safety monitoring can be extended to new lines or layouts without rebuilding analytics, according to the source material.
Edge system stops machines and triggers alerts when cameras detect unsafe conditions
AI cameras using object detection and tracking modules flag unsafe conditions in real time. deviceWISE combines machine data, robot status and process context to judge severity, then acts at the edge. Actions include sending a targeted alert, stopping a machine or orchestrating safety workflows. The system also captures structured data for reporting and continuous improvement.
The integration targets manufacturers seeking shorter detection-to-action times and consistent product quality across lines.