The Burden of
Server Infrastructure
Replacing a surveillance system shouldn't mean building an entire server infrastructure from scratch.
The Real Barriers to VLM Surveillance Adoption
The Burden of
CCTV Replacement
Costs
There is simply no budget to tear down and replace all the existing CCTVs already in place.
Overwhelmingly Complex Onboarding
Even running a simple PoC feels daunting because the evaluation and adoption process is so complicated.
Enterprise-Centric
Solutions
It is hard to find a tailored product for small-scale sites that only need 4 to 8 channels.
Uncompromised Core Features of
High-Performance VLM
Pre-loaded Scenario-Specific VLM Models
Deploy pre-loaded VLM models to start high-performance monitoring with zero training.
Built-in VMS & NVR
Start monitoring instantly without extra software and store up to 3 months of footage.
Natural Language Video Search
Type what you are looking for to instantly find the exact scene.
API / SDK Support
Seamlessly integrates with existing systems like MES, PLC, and access control.
Built Exactly for These Workplaces
Sites seeking high-performance AI without extra infrastructure costs.
Small-Scale Manufacturing Plants
- Detecting missing helmets & PPE compliance
- Danger zone alerts for forklift radiuses & restricted areas
Sites requiring targeted surveillance without replacing the whole system.
Unmanned Stores & Retail
- Detecting after-hours break-ins & loitering
- Instant alerts for late-night collapses, vandalism & disruptions
Logistics Warehouses
- Early smoke & fire detection in blind spots
- Real-time control of unauthorized access to security zones
Sites starting small with a flexible 4-to-8 channel setup.
Construction Sites
- Worker proximity alerts near heavy machinery (excavators, etc.)
- Post-work intrusion prevention in temporary construction zones
Apartments &
Facility Management
- Detecting loitering & abnormal behavior in garages
- Alerts for trespassers & elderly collapses in the complex
Sites testing high-performance VLM before launching a major rollout.
Schools &
Educational Facilities
- Detecting after-hours & weekend school trespassing
- Real-time detection of school violence & abnormal behavior
Connect your way, tailored to your site.
Flexible integration: Standalone, VMS/NVR, and Robotics/IoT.
Small-Scale Standalone Monitoring
Direct RTSP Camera Connection
Built-in NVR for Local Video Storage
Compare details, simplify choices.
CCTV Compatibility
Server Setup Required
Min. Deployment Scale
VLM Support
Deployment Cost
Odin AI Edge
Use Legacy CCTV
No Server Needed
From Min. 4ch
Supported
Low Cost
Company A AI Box
Separate Purchase Required
No Required
From Min. 32ch
Not Supported
Moderate
Company B AI Box
Separate Purchase Required
Required
From Min. 32ch
Conditional Support
High
Hardware Specifications
Item
Specifications
GPU
NVIDIA Jetson Orin Nano 8GB (67 TOPS) or NVIDIA Jetson Orin NX 16GB (up to 100 TOPS)
Memory
LPDDR5 (AI models preloaded, up to 16GB)
Storage
M.2 2280 NVMe SSD, up to 512GB (configurable upon request)
Network
2 × Gigabit Ethernet (GbE) LAN — RTSP/ONVIF camera and NVR integration
Wireless (Optional)
Wi-Fi + Bluetooth (M.2 2230), 4G/5G module (M.2 3052/3042, dual nano-SIM) — sales configuration to be confirmed
Video Output
HDMI 2.0b (4K/60Hz)
Expansion I/O
USB 3.2 Gen1 Type-A, COM (RS-232/422/485), DIO
Power Consumption
30W or less
Cooling
Fanless
Operating Environment
Temperature: −20–50°C / Humidity: 10–95%
Durability & Security
MIL-STD-810H vibration and shock compliance, TPM 2.0
Dimensions & Weight
152 × 114 × 72mm (approx. 1.25L) / approx. 1.4kg
FAQ
How many channels can one device analyze?
While it depends on your installation environment and video settings, Odin AI Edge is optimized for small-scale sites with 4 to 8 channels.
Which CCTV cameras are compatible?
It connects with most IP cameras supporting standard protocols like RTSP and ONVIF. Since it is brand-agnostic, you can seamlessly leverage your existing CCTV infrastructure without being locked into a specific manufacturer.
Can it be used with existing VMS/NVR?
Yes. You can add AI analytics capabilities while fully maintaining your current VMS and NVR setup. Note: Compatibility may vary depending on your specific system environment.
How do we scale up if the number of channels increases?
You can easily scale up by adding more Odin AI Edge devices depending on your site's scale. The expansion architecture and licensing policies are customizable based on the deployment size.
Is video data transmitted externally?
No. All video analytics are performed entirely on-device, and raw video footage is never transmitted to external servers. This minimizes data leakage risks and ensures high-level security.
Are there any additional SW or licensing fees?
No, the core software costs are fully included in the hardware price. With built-in VMS and NVR capabilities, you can use the system without any separate license purchases or subscription fees.
How long can video footage be stored?
Retention periods vary depending on the number of channels, video resolution, and bitrate settings. Under standard operating conditions, it supports up to approximately 3 months of video storage.
What are the hardware specifications?
Odin AI Edge features an AI SOM powered by the NVIDIA Jetson Orin Nano 8GB (67 TOPS), supporting up to 4 channels of analytics. It provides 2 default ports (4-port option available) with built-in PoE and VMS/NVR functionalities, all operating within a low-power design under 30W.