# Onyxia is Still Under the Radar

> Onyxia is a platform that can be installed on-premises, providing self-service technologies to data scientists. I find it very interesting, it should be promoted more!

- URL : https://blog.xreveillon.eu/en-US/blog/onyxia-outsider
- Date : 2025-09-14T08:03:54.000Z
- Langue : en-US

---

## What is Onyxia?

[Onyxia](https://www.onyxia.sh) is an open-source platform that enables deployment of work environments for data scientists. It's designed to be installed on-premises, providing full control over data and computing resources. The platform simplifies the management of development environments, libraries, and dependencies, allowing data scientists to focus on their work without worrying about technical configuration.

When connecting to the platform, users can choose from a service catalog which technology they want to install and use. For example, a JupyterLab with specific machine learning libraries, or RStudio with statistical analysis packages. The platform automatically handles the creation and configuration of these environments, which greatly simplifies the process for users.

<figure>
  <a href="/images/blog/onyxia-outsider/onyxia-datalab-services.png"><img src="/images/blog/onyxia-outsider/onyxia-datalab-services.png" alt="List of available services in Onyxia" /></a>
  <figcaption>List of available services in Onyxia</figcaption>
</figure>

## Open-Source Software from France!

Moreover, it's software developed in France, initiated by [INSEE](https://www.insee.fr/fr/accueil)! Companies are already using it, such as [Statistics Norway](https://www.ssb.no/en).

## How to Install?

Everything is explained in the [official documentation](https://docs.onyxia.sh/), with several step-by-step guides for key stages.

<figure>
  <a href="/images/blog/onyxia-outsider/onyxia-datalab-installation-details.png"><img src="/images/blog/onyxia-outsider/onyxia-datalab-installation-details.png" alt="Onyxia installation details" /></a>
  <figcaption>Onyxia installation details</figcaption>
</figure>
