SAI-HUB VID
Desktop software application for professional object detection from any video stream. Turn visual data into actionable intelligence.
Mission Profiles
Missions act as the command center for your detection logic. Store and recall specific configurations for different deployment scenarios.
- Algorithm Locking: Save specific models to specific streams.
- Object Definition: Target only the classes that matter.
- Confidence Thresholds: Eliminate false positives with precision tuning.
Algorithms & Neural Networks
Select the brain of your operation. SAI-HUB VID supports a wide range of computer vision models suited for any industry.
Pre-Loaded Models
Standard libraries for immediate deployment across common objects.
Custom Training
We can train bespoke models based on your specific visual data requirements.
BYOM (Bring Your Own Model)
Directly upload your own .pt files for custom proprietary detection.
Core Capabilities
Advanced features designed for high-precision, low-latency deployment environments. Turn incoming video data into real-time metrics.
Multi-Streaming
Simultaneous detection on 4 different camera streams.
Target Object
Infrastructure specific filtering.
Active Alarm
Real-time triggered alerts and SMS notifications via integrated SIM.
Video Archive
Automated clip peak extraction.
Brain-Box Storage
Every detection is an asset. SAI-HUB VID captures comprehensive metadata for every frame, stored exclusively on-premises on the Brain-Box for total security.
{
"Date": "2026-02-20",
"Latitude": "48.2345",
"Longitude": "2.5423",
"Number of Objects": "3",
"Project": "Secure_Facility_Monitoring",
"Targeted Object": "Vehicle",
"Confidence": "87%",
"Inferenced stream": "rtsp://:xxxx/"
}
Hardware-Locked Storage
Data never leaves the Brain-Box. Logs are stored in structured JSON and text format for easy export to existing VMS or analytics tools.
Spatial Intelligence
Includes GPS coordinates and precise bounding box metadata for every detection, allowing for accurate geospatial mapping.
Stream Inference
Real-time processing across multiple RTSP/ONVIF streams simultaneously, handled locally by the powerful Brain-Box engine.