Metrici HighSYS-AI Server

Product Overview — High-Performance AI Server for Vision Analytics

Welcome! This is a short overview of the Metrici HighSYS-AI — the high-performance tier of the Metrici AI Server family. It is a pre-configured, on-premise appliance that ships with Ubuntu and the Metrici detection engines already installed. This page covers the key specifications, where the model fits in the lineup, what it runs, and how to bring it online.

Introduction
Technical Data
What It Runs
Setup & Operation
Support & Resources

1. Introduction

The Metrici HighSYS-AI is a high-performance, workstation-class AI server built to run several Metrici deep neural network engines at the same time. With a discrete Nvidia GPU working alongside the Intel NPU, it is made for demanding multi-camera environments — running LPR, parking detection, and thermal analysis simultaneously with headroom to spare.

The unit is delivered as a ready-to-use appliance: the operating system and Metrici engines are pre-installed, so no software setup is required on site. The engines are then customised for the specific deployment — adding cameras, configuring detection parameters, and completing setup in the Metrici Web Interface. All processing happens locally on the unit; no data has to leave your premises.

At a Glance

Processor

Intel® Core™ Ultra 5 245K · 4.2 GHz

Graphics (GPU)

Nvidia RTX 5060 Ti · 8 GB

AI Acceleration

Nvidia Tensor Cores + Intel® AI Boost NPU

Camera Capacity

Up to 10 cameras

Memory / Storage

16 GB DDR5 · 500 GB SSD

Operating System

Ubuntu 24.04 LTS

The Server Range

Metrici AI Servers come in six tiers, from a compact edge device to a data-center-scale platform. HighSYS-AI sits in the high-performance tier — the choice for multi-engine deployments and campus-wide or industrial camera coverage that benefits from GPU acceleration.

MicroSYS

→

MidSYS

→

HighSYS

→

UltraSYS

Each tier is pre-validated with Metrici software before delivery, and the architecture is designed to scale — you can move up to a higher tier as a deployment grows. If you are unsure which model fits, Metrici’s team sizes the system based on your cameras, engines, and expected workload.

2. Technical Data

The full specification of the HighSYS-AI as delivered:

Specification	Details
Architecture	Intel x86_64
Processor (CPU)	1 × Intel® Core™ Ultra 5 245K, 4.2 GHz
Processing units	14 cores (6 P-cores + 8 E-cores) / 14 threads
GPU	1 × Nvidia RTX 5060 Ti, 8 GB
AI accelerator	Nvidia Tensor Cores (GPU) + Intel® AI Boost NPU, 13 TOPS
Performance	Up to 10 cameras, depending on workload type
RAM memory	16 GB DDR5
Storage	500 GB SSD
Networking	1 × Ethernet 10 / 100 / 1000 Mbps
Graphics	Discrete, HDMI / DisplayPort output
Format	Cube case, 34 × 32 × 43 cm
Power source	1 × 850 W
Operating system	Linux Ubuntu 24.04 LTS

Note: AI acceleration on the HighSYS-AI combines the discrete Nvidia RTX 5060 Ti (Tensor Cores) with the integrated Intel® AI Boost NPU — the GPU is what gives this tier the headroom to run several heavy engines at once. Sustained camera throughput still depends on the mix of engines running; see the next chapter.

3. What It Runs

The HighSYS-AI is configured to run any combination of Metrici detection engines at the same time, and the discrete GPU lets it handle several heavy engines together. Common engines include:

Metrici LPR — licence-plate recognition with multi-country support and vehicle classification.
Metrici PPD — parking-place detection for garages, structures, and open lots.
Metrici Thermal Analyzer — AI thermal analysis for monitoring and predictive maintenance.
Metrici Area / Line Counter — people and object counting.
QR Code & Container Code recognition, and other custom AI workloads.

The unit supports up to 10 cameras. The exact number it can drive comfortably depends on the engine mix, frame rates, and detection load — a single light engine reaches the upper figure more easily than several heavy engines running together.

Sizing tip: If you are planning several engines on the same unit, share your camera count, engine list, and expected traffic with Metrici. The team will confirm whether HighSYS-AI is the right tier or recommend a step up.

4. Setup & Operation

The HighSYS-AI arrives ready to use. To bring it online for the first time:

Place the unit on a clean, flat, well-ventilated surface (see Placement).
Connect the network cable to the Ethernet port.
Connect a display to the HDMI or DisplayPort output if needed. The unit can also run headless.
Connect the unit to mains power using the supplied power cable, then switch on the power supply.
Power the unit on.

The operating system completes boot within roughly 30–60 seconds, and the Metrici engines start automatically as background services. The unit’s IP address is either pre-configured by Metrici before shipping (documented in the delivery note) or obtained from your network by DHCP. From there, configuration is completed in the Metrici Web Interface.

Placement & Ventilation

Install the unit on a stable surface in a well-ventilated area, away from direct sunlight, heat sources, and high-dust environments. Leave clearance around the chassis so air can move freely through the vents, and avoid enclosed cabinets without active ventilation. Because this tier includes a discrete GPU, it produces more heat under load than the lower tiers, so unobstructed airflow matters even more.

Important! Do not block the chassis vents. Operating the server with restricted airflow will cause the processor and GPU to thermally throttle — reducing performance — and may shorten product lifetime.

5. Support & Resources

For configuration help, system sizing, or a quote for the right server tier, contact the Metrici team. The product datasheet and CE declaration are linked in the sidebar under Resources.

Web: www.metrici.ro
Sales: sales@metrici.ro
Support: support@metrici.ro