• This article describes how you can start an instance with JupyterLab preinstalled and how to access it. 

  • For this tutorial, we assume that you are already able to configure and run an instance, if not you can find a guide on how to start an instance in our Knowledge Base articles. 

  • If you prefer to install JupyterLab with your instance, you can refer to this guide.


Configuring your instance


  • When configuring a new instance, you should select 'JupyterLab' under the 'Catalog' tab of the 'Select Image' section. After selection, choose the desired 'Base Image' for the instance. 

  • Finally, you can generate or input the 'Web Dashboard Password' of your choice which is required to access the JupyterLab dashboard. 

  • Please Note: this password is only used when authenticating in the Web Dashboard and different from the authentication method you set up for accessing the terminal of your instance. You must copy the password at this point & save it at some safe location, because it will not be made available after the instance creation. 

  • If you don’t want to take care of the GPU driver installation yourself, make sure to select the 'Install NVIDIA® GPU driver 470' checkbox.




Accessing your Jupyter Dashboard


  • After configuring and creating your instance with JupyterLab, you should wait for 10-15 minutes before accessing the instance, so all the required packages get installed. Then, head over to the instances list and click on your newly created instance. Here you will find the URL for the Jupyter Dashboard under 'Catalog Details'.


  • Click on that URL, new Jupyter Dashboard window will open, you will be prompted for the password you set when configuring the instance.


  • After entering the password, you will be able to access the Jupyter Dashboard. Now, you are ready to start working with your selected Jupyter interface.



  • Remember to stop or destroy your instance when you are done using the JupyterLab. A stopped instance can be started again later. Closing the dashboard does not shut down your instance.