Edge connection hardware that enables to running of deep learning solutions on IP cameras.

 Preorder now 


Visionair creates powerful hardware links between IP cameras and deep learning software to open doors to new opportunities for AI solutions.

The device supports a connection to H.264 type of video format and the capability to plug in up to 4 cameras.    

This hardware gives the flexibility to choose from NVIDIA 21 TOPs or 32 TOPs Jetson GPUs and adjust the performance power according to your needs. Designed with the environment in mind, Visionair has low energy consumption with an option to power IP cameras and the hardware from solar panels. The device supports 4G/5G connection and seamlessly integrates into the surroundings. High productivity in three simple steps: install, connect and deploy!    

Boost productivity of IP cameras with real deep learning

Visionair offers an opportunity to optimise resources by automating operational processes with object detection technology.

Visionair has powerful NVIDIA GPUs that provide a perfect environment to execute software of any complexity.

A unique cooling system is designed to maintain a stable temperature and secure GPU from overheating. 

Tech Specs

M ax. 4 H264 streams (depending on the pipeline)  

         NVIDIA Maxwell™ architecture with 128 NVIDIA CUDA® cores 0.5 TFLOPs (FP16)  

  • NVIDIA TX2 NX   
         NVIDIA Pascal™ Architecture GPU with 256 CUDA cores, 1.33TFLOPS  

         384-core NVIDIA Volta™ GPU with 48 Tensor Cores; 21 TOPs  

4G, 5G, Ethernet (default)

1 year

-10°C to + 60°C

Programming compatibilities

Open source; Python

Watch the interview to learn all details about Visionair!


We create ultimate object detection hardware and awesome content! 

Share your email to stay updated

Protected by reCAPTCHA, Privacy Policy & Terms of Service apply.