Train, test, and deploy AI in Unity. Generate synthetic data, run ONNX models, and connect to real PLCs with realvirtual.io AI Builder for industrial vision and automation.Train, test, and deploy your AI – directly inside Unity.realvirtual.io AI Builder turns Unity into a full-featured platform for synthetic data generation, AI model testing, and industrial deployment. Whether you're building machine vision for robotics, logistics, quality inspection, or production lines, AI Builder lets you train smarter AI using Digital Twins.✅ Key FeaturesCreate rich synthetic datasets with labeled images, depth maps, segmentation masks, and 2D/3D bounding boxesIntegrate ONNX models for oriented object detection, segmentation, and classificationUse domain randomization and real-time physics to train more robust AIRun & test YOLO and other ONNX models directly inside your Unity sceneGenerate training data at scale using realvirtual.io’s automation frameworkDeploy AI across platforms – Windows, Linux, Android, macOSConnect to industrial automation: Siemens S7, Beckhoff ADS, OPC UA, MQTT, Modbus, WebSocket (Industrial interfaces available in realvirtual.io Starter or Professional editions)• Compatible with Siemens S7, Beckhoff ADS, OPC UA, MQTT, Modbus, WebSocket - Industrial Interfaces available in realvirtual.io Starter or realvirtual.io Professional🧪 YOLO Training IntegrationIncludes a ready-to-use example pipeline for Ultralytics YOLO with synthetic data generation and model export workflow.We provide two open-source training pipelines (AGPL-3.0 license):Note: Ultralytics is not included in this asset. It can be downloaded as open source under the AGPL-3.0 License from:https://github.com/ultralytics/ultralyticsWe provide two open-source training pipeline Examples under AGPL-3.0 License:• A Toy Brick Detection pipeline• A Card Detection pipeline ( the full Igus Rebel UI Application is not included).These examples can be used as templates to build your own ONNX-compatible models with YOLO. Other pipelines are supported but may require custom scripting.🚀 Designed for:Vision system developersRobotics and automation engineersIndustrial AI prototypingSmart factories & logistics systemsAcademic & research simulation labs🎮 Built on top of realvirtual.ioAI Builder extends the realvirtual.io framework – the industrial Unity toolkit for real-time physics, PLC connectivity, and Digital Twins.💡 BonusIncludes demo scenes and a YOLO ONNX test example for immediate productivitySeamless integration with Unity Editor for rapid prototyping and testingNo cloud required – all workflows run entirely offline if needed📸 Synthetic Data GenerationGenerate high-quality labeled images and bounding boxes inside Unity.🧠 ONNX Model IntegrationLoad and run ONNX models (e.g., YOLO) directly in your scene using Unity Sentis or third-party runtimes.🎯 Domain RandomizationAutomatically vary lighting, textures, object positions, and camera angles to improve AI generalization.🔍 AI Testing in Digital TwinsSimulate, visualize, and validate AI behavior inside a working Digital Twin – before deployment.📦 Scalable Data PipelinesAutomate large-scale dataset creation with controlled scene variation and labeling tools.🔗 Real-Time Deployment Connect your AI to real PLC systems via Beckhoff ADS, Siemens S7, OPC UA, MQTT, Modbus, or WebSocket - requires realvirtual.io Starter or Professional industrial interfaces🌍 Cross-Platform SupportBuild and run AI-enabled Unity apps on Windows, Linux, macOS, Android, and more.🛠 Industrial-Grade DemoIncludes a ready-to-run Lego and Card sortingexample with real-time classification and sorting logic.🧩 Compatible with realvirtual.ioSeamlessly integrates with the realvirtual.io framework for physics-based automation and Digital Twin development.
