AI-based application software for product-quality inspection during production phase
  • High-precision diverse algorithms
  • Zero-code model generation
  • Model validator: Adjust model sensitivity
  • Model predictor: Predict with test data
  • Windows 10 and Windows 11 64-bit editions supported
  • English, traditional Chinese and simplified Chinese languages supported
  • Notification and software-upgrade mechanism
 A male operator testing the product with a display screen showing a software interface in front of him.


ASUS IoT AISDetector is an on-premise AI signal-analysis and anomaly-detection tool empowered by AI technology developed by ASUS IoT. AISDetector facilitates the process by providing both training and inference working models, thereby simplifying AI-model development and management. Developers can utilize AISDetector for building specialized AI models, tracking training progress information, performing model training, model validation and conducting anomaly-detection testing.​

AISDetector provides a user-friendly interface for model training and validation, and then leverages resulting models to test for defective products. The system includes a versatile web API to enable developers to draw signal-data inferences derived from a variety of sensors, including vibrations, sound and current – all powered by a unique and ever-evolving ASUS IoT AI algorithm.​

AI Software Project Management

Managing multiple AI models and use cases

Utilize AI project tools to manage essential information for various signal scenarios. Develop anomaly-signal-detection models effortlessly using the AISDetector zero-code AI toolkit.
The image depicts the interface for AI software project management.

Multi-Sensor Data Integration

Supporting common data formats for signal acquisition or import

Our system seamlessly integrates with three common signal data import formats: CSV, JSON and WAV. It supports data sources from sensors such as vibration, sound, voltage or current. Quality signal data enables efficient model training through AISDetector.
The image depicts the interface for multi-sensor data integration.

One-Minute Model Training

Rapid model development

Train an effective AI model in just one minute, and with just four simple steps. This rapid training process allows for quick establishment of models to detect anomaly signals. Unsupervised learning requires only CPU for training and verification.
The image depicts the interface for one-minute model training.

Intuitive Model Validation

Facilitates quick model validation and tuning

Easily validate and fine-tune the model by adjusting its five key features. Once AISDetector has generated and validated a model, it can be exported for further application development through AISDetector's extensive API libraries (C, C++, C#, Python).
The image depicts the interface for intuitive model validation.

Built-in Model Testing Tools

Convenient Verification Tool

AISDetectror makes it easy to conduct tests with real-time pass/fail results displayed on the screen through the inference of live signal inputs.
The image depicts the interface for built-in model testing tools.


  • Manufacturing


    Defect Detection

comparison icon
Compare (0)