Minimum (up to 5 cameras @ 15–20 FPS)
| CPU | 4+ cores — Intel i5 / AMD Ryzen 5 |
| RAM | 16 GB |
| Storage | 256 GB SSD |
| GPU | Not required |
| OS | Ubuntu 20.04/22.04, Debian 11+, Windows 10/11 with WSL2 |
| Network | Gigabit LAN for camera streams |
Requirements
Runs on standard hardware — no GPU required
CPU-optimized AI means you deploy on existing infrastructure without expensive GPU upgrades.
Recommended (10+ cameras @ 25–30 FPS)
| CPU | 8+ cores — Intel i7/i9 / AMD Ryzen 7/9 |
| RAM | 32 GB (64 GB for heavy setups) |
| Storage | 500 GB+ NVMe SSD |
| GPU | Not required |
| OS | Ubuntu 22.04 LTS Server (recommended) |
| Network | Gigabit Ethernet, managed switch |
Performance
Expected throughput
| Hardware | Cameras | Face Recognition | Concurrent Users |
|---|---|---|---|
| Minimum (4 core, 16 GB) | Up to 5 @ 15–20 FPS | 2–3 seconds | 5–10 users |
| Recommended (8+ core, 32 GB) | Up to 10+ @ 25–30 FPS | < 1 second | 20+ users |
Cameras
Supported camera brands
FAQ
Frequently asked questions
Do we need a GPU?
No. The system uses CPU-optimized AI processing. Runs on standard Intel i5/i7 or AMD Ryzen hardware. No dedicated graphics hardware, no cloud AI API, no additional hardware cost.
How many cameras can it handle?
Minimum hardware (4-core, 16 GB): ~5 cameras at 15–20 FPS. Recommended hardware (8+ core, 32 GB): 10+ cameras at 25–30 FPS.
What camera brands are supported?
Any camera with IP camera or HTTP stream. Tested with Dahua and Hikvision. Any standard IP camera camera works.
Ready to make your cameras intelligent?
Request a demo, download the datasheet, or talk to our team about on-premise deployment for your facility.