Skip to main content
Smart Visual Hybrid
SVH - Smart Visual Hybrid

Intelligent industrial
XR training

Extended reality (XR) with artificial intelligence for industrial operators. Immersive training, smart assistance and plant monitoring. No cloud.

View modules

Four modules. One ecosystem.

Functional MVP

Training XR

Immersive extended reality (XR) training for industrial operators. Step-by-step procedures with visual guidance on XR headsets.

OpenXROpenUSDUnity
Prototype

Smart Recognition Assistant - ARI

Mixed reality (MR) powered by AI for operator assistance. Visual recognition of parts and tools, workstation verification and real-time voice guidance.

Computer VisionWebRTCEdge AI
Functional MVP

Inmersive Space

Real-time industrial monitoring with sensor data overlaid in the operator's real environment.

OPC-UAIoTDigital Twin
Active service

Custom development

Sector-specific XR/AI projects. SVH platform adaptation to each industrial client's specific needs.

Custom XRSectorialB2B

Frequently asked questions

What is SVH (Smart Visual Hybrid)?

SVH is an extended reality (XR) and mixed reality (MR) platform with embedded computer vision, designed for industrial operators. It combines immersive training, AI-powered smart assistance and plant monitoring on the same technical architecture.

Does it work without a cloud connection?

Yes. SVH runs all AI inference and computer vision processing locally via edge computing. Sensitive data never leaves the plant, ensuring data sovereignty and operation even without internet connectivity.

How does SVH learn to recognize new parts?

The ARI module uses one-shot learning: from a single visual sample the system identifies parts or tools without lengthy retraining. This dramatically reduces deployment time at new workstations.

How does SVH differ from enterprise XR platforms?

SVH combines XR + AI + edge computing in a single product, with accessible hardware and no cloud dependencies. Traditional enterprise platforms require costly licenses, cloud infrastructure and usually do not integrate computer vision or one-shot learning.