Contents
System Requirements
Minimum and recommended hardware, storage, network, and browser requirements for Kalytera VMS.
Minimum Hardware
These are the minimum specifications required to run Kalytera VMS:
| Component | Minimum | Recommended |
|---|---|---|
| CPU | 4-core (Intel/AMD x64) | 8-core modern CPU (Intel i7/Xeon, AMD Ryzen 7/EPYC) |
| RAM | 4 GB | 16 GB (32 GB for large deployments) |
| Storage (OS + App) | 20 GB SSD | SSD for OS and app; separate HDD/NAS for recordings |
| Network | 100 Mbps Ethernet | Gigabit Ethernet; cameras on a dedicated VLAN |
| Operating System | Windows 10 / Windows Server 2016 | Windows Server 2022 / Ubuntu 22.04 LTS |
For deployments with 16+ cameras or AI analytics enabled, use the recommended (or higher) specifications. CPU usage grows primarily with the number of cameras using AI object detection.
Storage Calculator
Recording storage requirements depend on resolution, frame rate, bit rate, and retention period. Use this as a rough guide:
| Resolution | Typical Bit Rate | GB per camera per day |
|---|---|---|
| 1080p (Full HD) | 2–4 Mbps | 21–43 GB |
| 4MP | 4–6 Mbps | 43–65 GB |
| 4K (8MP) | 8–12 Mbps | 86–130 GB |
| 720p (HD) | 1–2 Mbps | 11–22 GB |
Kalytera uses H.264 passthrough recording (no re-encoding), so storage usage matches the camera's configured bit rate exactly. Lower the camera's bit rate in its own web interface to reduce storage consumption.
Network Requirements
- Camera bandwidth: Each camera's stream bit rate × number of cameras. Plan for at least 1.5× headroom.
- Viewer bandwidth: Each live viewer adds another stream's worth of bandwidth from the server. WebRTC direct paths reduce server load.
- Recommended: Place cameras on a dedicated VLAN, isolated from general office traffic.
- Remote access: Port-forward
5099on your router, or use a VPN for secure remote access. Avoid exposing the server directly to the internet without proper firewall rules.
GPU for AI Analytics
Kalytera's AI object detection uses OpenVINO (Intel) or ONNX Runtime (CPU fallback). A dedicated GPU is not required, but improves throughput with many cameras:
- Intel CPU / iGPU: OpenVINO runs efficiently on Intel CPUs and integrated GPUs via the Intel HD/Iris/Arc driver stack. No additional software required.
- NVIDIA GPU: Not natively supported via OpenVINO in the default build. Use the ONNX Runtime CPU path as a fallback.
- CPU-only: Adequate for up to 8–12 cameras at 4 FPS analytics without a GPU.
Browser Support
Kalytera's web interface requires a modern browser with WebRTC support:
| Browser | Support Level |
|---|---|
| Google Chrome 90+ | Full (recommended) |
| Microsoft Edge 90+ | Full |
| Firefox 90+ | Full |
| Safari 15+ | Full (WebRTC may require HLS fallback on some camera profiles) |
Supported Camera Brands
Kalytera supports cameras via native protocols and ONVIF:
- Hikvision — native SDK + ONVIF + RTSP
- Dahua — ONVIF + RTSP (AUD filter applied automatically)
- Axis — native ONVIF + RTSP with WebRTC optimization
- Bosch — ONVIF + native RTSP stream override
- Generic ONVIF — any ONVIF-compliant IP camera
- Generic RTSP — any camera exposing an RTSP stream