Understand more. Build with confidence.
Official guides and references to plan, build, and run Vision AI.
2D vs. 3D Vision: When Does Inspection Need Height and Depth Data?
02Visual AI for Wood and Furniture: Knots, Cracks, and Finish Defects
03AQL Sampling vs. 100% Inspection: When Should AI Cameras Take Over?
04AI Cameras for Ceramics and Glass: Cracks, Chips, and Bubbles
05Smart Inspection for Agricultural Sorting: Coffee Beans, Cocoa, and Fruit
06Machine Vision for Metal Surfaces: Scratches, Dents, and Rust
Do I need to know how to code?
Nirmana’s core workflow is visual: projects, datasets, annotation, training, inspection, rules, and integration. Camera setup and systems integration still require process knowledge and technical validation.
Which cameras can I use?
Documented sources include USB/webcams and IP cameras/CCTV with RTSP. Industrial cameras require verification of drivers and stream access; do not assume universal compatibility.
Does data have to be sent to the cloud?
Nirmana supports local/on-premise processing. Packages are available for Windows x64, Linux x86_64, and NVIDIA Jetson aarch64. Internet is required for activation and periodic license verification.
How is free mode different from a trial?
Without a license, Nirmana runs in a limited mode with one camera. Full-feature trials require an approved request and are time-limited. Trial activation is not automatic.
Can I connect PLC, MES, or ERP systems?
Documented integrations are Modbus TCP and HTTP/webhooks. All editions include PLC/API integration. Register mapping, payloads, and system workflows need engineering validation.
How accurate and fast is it?
Results depend on datasets, models, resolution, lighting, camera count, and hardware. Test representative data to establish feasibility. Website demo numbers are simulated, not product specifications.
Your camera can do more.
Build your Vision AI. Connect it to the real world.