Cordatus AI: Deploy AI Environments, Applications and Manage Containers on NVIDIA Jetson

Hi all,

After two years of hard work, we are happy to announce our new software; Cordatus AI.

Cordatus is an AI platform that supports NVIDIA Jetson Nano/TX2/Xavier NX/AGX Xavier and NVIDIA GPU enabled PCs, Workstations, Servers. The platform has lot to offer for NVIDIA Jetson users. It features:

  • Optimized AI environment containers for NVIDIA Jetson (TensorFlow, PyTorch, MXNet, OpenCV, DLib and more).
  • Optimized and ready-to-run AI application containers for NVIDIA Jetson (Detectron2, TensorRT Pose Estimation, YOLO v3, Tensorflow Object Detection API, TF-TRT and more).
  • Public URLs for Jupyter Notebooks that runs within environment and application containers (Remote Jupyter Notebook Access).
  • Low-latency live stream from CSI, GMSL2, USB and IP cameras both locally and remotely.
  • Low-latency TensorRT optimized object detection across devices (process Jetson’s CSI stream on Workstation or AWS EC2 instance).
  • Detailed monitoring for all NVIDIA Jetson devices (CPU Usage, CPU Temperature, Shared Memory Usage, GPU Utilization and GPU Temperature).
  • Switching between Jetson’s performance modes from both Cordatus Web and Cordatus GUI/CLI App.
  • Enable/disable jetson-clocks.
  • Remotely restart or shutdown your Jetson devices at the edge.
  • Deploy, stop, start or delete your remote AI environment/application containers remotely using Cordatus Web.

You can check out our blog post series on Medium and create an account for free to start experimenting with Cordatus AI.


1 Like

That looks like a good project. Since much of the Jetson support environment uses an Ubuntu 18.04 host PC, is this also intended to run on Ubuntu 18.04? Or will this run well on different distributions due to the containers?

Hi, thanks for your reply. It is widely tested and runs perfectly on Ubuntu 18.04 but we are currently testing it on Ubuntu 20.04 as well.

Once we are done testing with Ubuntu since it is one of the most widely used distributions, we will be working on other Linux distributions too.

Hi @doruk.snmz
Can you explain the limitations of the free plan/standard tier a little bit more in detail?

What is the intended use-case/audience of the different plans, e.g. I am hobbyist user/open source developer and own a Nano, a Xavier-NX and a Xavier-AGX - will the free plan work for me?

Hi @dkreutz. Thanks for your comment.

Free plan could be useful for most of the use cases for hobbyists. However if you planning to experiment with multiple camera streams, you may consider other plans.

You can check out our new blog post that describes the features in scope of licensing in detail from the below link;