Thanks! Will do some more user testing with the navigation to see if some things can be ironed out.
The main API runs fully in the browser and is JS based with JSX for visualization and displaying components like charts, lists, tables, etc. By using JS it also keeps it accessible for a wide audience to quickly put something together.
The ML/AI plugins are called from the JS api and run on a VM that currently only supports Python, with the usual libraries such as scikit-learn, etc. In the near future we'll add more languages and libraries.