System requirements
Metrici video-analytics software · self-hosted PC or server
Metrici runs as a set of modular AI engines — LPR, Area Counter, Line Counter, Parking Place Detector and others — each handling one recognition task. Every active engine adds CPU, memory and disk load, and that load also grows with the number of cameras, their resolution and frame rate.
The configuration below is sized for a typical deployment running 1 to 4 engines. Beyond four engines, or for high-resolution multi-camera sites, scale up the GPU, cores, RAM and storage accordingly — or ask Metrici for a sizing review.
- Operating system
- Linux Ubuntu 24.04 LTS MATE
- Processor
- Intel 12th-gen Core i5 | AMD Ryzen 5 9000 series or above
- Memory
- 16 GB RAM
- Graphics
- NVIDIA GeForce RTX 2060 | AMD RX 9060 XT, or Integrated Graphics
- Storage
- 250 GB SSD — scales with data retention time (see notes)
- Networking
- Gigabit Ethernet 10/100/1000 Mbps
Storage scales with data retention time
Disk usage is driven by stored events, snapshots and recordings, so it depends entirely on the site. A test bench or low-traffic location runs comfortably on 250 GB; a busy site such as a mall car park can need 1 TB or more. Always use an SSD.
Reference appliance — MicroSYS-AI
The Metrici MicroSYS-AI is the smallest turnkey system Metrici ships: Intel Core Ultra 5 with AI Boost NPU (up to 34 TOPS), 16 GB DDR5, 500 GB SSD. The requirements above are the self-hosted equivalent for deploying on your own hardware.