Supported devices

With Scailable we make creating and managing highly performant edge AI solutions as easy as possible. To ensure the efficiency of our full edge AI pipline (i.e., from grabbing the camera images and pre-processing, to model inference and output post-processing) we focus on a carefully selected set of edge device. For each of the devices we support we provide extremely efficient, low-level integrations of our AI manager. This often includes automatically using the hardware acceleration (GPU, NPU, TPU) that is available on the device.
In this section of our documentation we list our supported devices and highlight were you can buy these devices with the Scailable AI manager pre installed. You can find links to quick dowloads of the AI manager here.

Supported devices


We intimately support a number of edge devices produced by Advantech:
  • Cellular routers. Our AI manager has been highly optimized for Advantech cellular routers. This means the Scailable AI manager can run highly efficiently on these routers, which in turn allows for the creation of mobile, stand-alone, edge AI solutions that include the desired networking in a single device. We suport:
    • The harware platform v3 routers. Here we explicitly support ICR-3xxx and the SmartFlex series. Note that these routers have limited processing and are thus mainly suited for relatively low throughput applictions.
    • The hardware platform v4 routers. Here we explicitly support the ICR-4xxx series. These routers are suprisingly flexible and can carry a large number of edge AI solutions.
    On the Advantech cellular routers the Scailabel AI manager is available as a router app.
  • The Edge AI box with NVIDIA Jetson (MIC-AIxx). Our AI manager runs highly efficiently on the full range of Advantech Edge AI box pc's with NVIDIA Jetson inside. Given the GPU acceleration on these machines these can scale up to extremely large and complex AI models that need to be executed at extremely high speeds.
Note that our AI manager comes pre-installed with a number of alternative edge devices (such as real-time IPCs and industrial camera's). In these cases our edge AI deployment solution is white-labelled. If you are interested in these applications please read out to our sales.


In collaboration with Siemens we provide support for the Scailable AI manager installed on the Siemens IoT 2050.

Linux based machines

Effectively, our AI manager can be installed on virtually any Linux based machine.
We currently do not provide support for edge devices other than the ones listed on this page.
Note that it is much easier to simply purchase our supported hardware with the Scailable AI manager preinstalled via one of our resellers. Learn more here.
Although you can find more information for each of the devices that we support on their respective pages, here we provide a quick overview table to retrieve downloads:
NVIDIA CUDA (aarch64) .tar.gz download
(although its easier to use our one-line install)
IoT 2050
Contact our support.
ARM 32 bits (armv7) ARM 64 bits (aarch64) Intel x86_64 (amd64) (although it is easier to user our one-line install)
Contact our support.
For all our supported devices we also have demonstration versions of our AI manager available; these demonstration versions contain easy on-device visualizations of the AI models for demonstration purposes. Learn more about setting up these demonstration versions here.