Track water quality and flow in real time for more efficient operations
WinEdge Edge AI inference system
Run AI models
in real time at every site
Deploy AI models near your data and devices.
Manage real-time inference across sites.
Inference latency< 8ms*
Data stays on-siteNo cloud
Multi-device managementCentralized
* Inference latency varies with model, hardware configuration, and application environment.
WinEdgeEDGE AI OPERATIONS
WHY WinEdge
Deploy Reliable AI at the Edge
Get low-latency results, connect equipment and manage devices at scale.
CAPABILITIES
Seven Edge AI Capabilities
Inference, Integration and Deployment
Run models, connect data sources and manage edge devices in one system.
AI Inference
Run vision AI inference in real time at the edge, with image classification, object detection and segmentation.
Analytics
Regression, classification, time-series and anomaly detection turn sensor and equipment data into results you can use right away.
Tracking
Set tracking targets with prompts or bounding boxes, count objects in ROI zones and trace paths to monitor people and objects.
Dashboards
Charts, SQL queries and filters, plus multi-source data integration, give you real-time visibility into inference and operating data.
Output Logic
Configure NMS, custom thresholds, logic rules and path tracking, with live display, data output and I/O triggers.
Deployment
A simple framework for deploying and updating models, with smart, centralized monitoring across devices.
Integrations
Multiple I/O connections, chained model workflows and custom developer modules extend WinEdge flexibly to any industry.
USE CASES
Edge AI Applications
Real-time on-premises inference for a wide range of edge use cases.
Water Monitoring
Track water quality and flow in real time for more efficient operations
AI spots defects in real time to raise process quality
Analyze traffic flow in real time to optimize traffic management
AI detects unusual situations in real time to keep spaces safe
FEATURES
Manage AI Across Devices
Connect data, deploy models and monitor devices in one place.
Smart Dashboards
Arrange live charts and tables in a drag-and-drop dashboard.
Connect MySQL, MSSQL, PostgreSQL, InfluxDB, Modbus TCP and industrial cameras.
Deploy, update and configure AIWinOps and NumOps models centrally.
Track device health, connections and resource use in real time.
Built for edge computing, WinEdge runs reliably without cloud resources or internet access, keeping on-site inference fast and data under your control.
WORKFLOW
How WinEdge Works
Connect data, deploy a model and monitor live results.
- 01Set Up Data SourcesConfigure data sources (MySQL, Modbus TCP, etc.) or imaging devices (industrial cameras, IP cameras) and pass the connection test.Set Up Data Sources
- Deploy Models Fast
- Set Up Inference Logic
- Deploy to Edge Devices
- Build Dashboards
- Device Monitoring and Management
INTEGRATION
Connect AI to Your Operation
Field equipment/IoT × WinEdge × WinHub.AI platform
Link equipment and systems to on-site vision and data models.
Equipment and IoT SensorsIndustrial and IP CamerasDatabases, Excel and CSVMES/ERP Systems
Data Integration
Equipment / IoT (Modbus)MySQL／PostgreSQL／InfluxDBExcel / CSV / Industrial CamerasMES/ERP Systems
Live Dashboards and Alerts
Multi-Chart VisualizationLive Status MonitoringAnomaly AlertsSQL Query / Filtering
Model Deployment
AIWinOps Vision ModelsNumOps Data ModelsVersions and ParametersRecipe Management
Inference Engine
Image Classification / Detection / SegmentationRegression, Classification, Time SeriesAnomaly DetectionObject Tracking
Post-Processing
Confidence FilteringI/O TriggersModbus Write-BackLive Display
Real-Time InferenceDevice Links and Control FeedbackMulti-Site ControlAutomatic UpdatesStatus Monitoring
Water MonitoringReal-Time Line InspectionTraffic ManagementSmart Childcare Safety MonitoringDevice Monitoring
01
Field Data and Equipment
Equipment and IoT Sensors
Industrial and IP Cameras
Databases, Excel and CSV
MES/ERP Systems
02
WinEdgeReal-Time Edge Inference Platform
Data Integration
- Equipment / IoT (Modbus)
- MySQL／PostgreSQL／InfluxDB
- Excel / CSV / Industrial Cameras
- MES/ERP Systems
Live Dashboards and Alerts
- Multi-Chart Visualization
- Live Status Monitoring
- Anomaly Alerts
- SQL Query / Filtering
Model Deployment
- AIWinOps Vision Models
- NumOps Data Models
- Versions and Parameters
- Recipe Management
Inference Engine
- Image Classification / Detection / Segmentation
- Regression, Classification, Time Series
- Anomaly Detection
- Object Tracking
Post-Processing
- Confidence Filtering
- I/O Triggers
- Modbus Write-Back
- Live Display
03
Edge DeploymentData Stays Local on a Plug-and-Play AI BOX
Real-Time Inference
Device Links and Control Feedback
Multi-Site Control
Automatic Updates
Status Monitoring
04
Applications
Water Monitoring
Real-Time Line Inspection
Traffic Management
Smart Childcare Safety Monitoring
Device Monitoring
DEPLOYMENT
Deployment Architecture and System Requirements
Deploy on-premises and keep your data and models under your control.
Software requirements
| Item | Version |
|---|---|
| Operating System | Ubuntu 22.04 LTS (64-bit) or later |
| NVIDIA GPU Driver | 580.x or later |
| Docker Engine - Community | 24.x or later |
Deployment architecture
- Flexible on-premises / private cloud deployment
- Data stays 100% on-site, under your control
- Multi-user collaboration and access control
- Hardware resource scheduling and quota management
Hardware requirements
| Item | Minimum | Recommended |
|---|---|---|
| CPU | 6 cores / 12 threads or more | Intel® Core™ i7 / server-grade Xeon® |
| GPU (optional) | VRAM 4GB or more, CUDA 7.5 | RTX 4080 / 4090 ～ NVIDIA A100 / H100 |
| Memory | 16GB | 32GB or more |
| Storage | 1TB SSD or more | Scale with workload |
| Ports | USB 2.0 / 3.0 (for security dongle) | |
Recommended compatible hardware
PRICING
Purchasing Options
Choose the licensing model that fits your use case and budget.
Perpetual Plan
- Perpetual license
- Private cloud / on-premises deployment
- Full control of your data
- Lifetime technical support
Best forLong-term use, sensitive data