Open source digital infrastructure

The open-source infrastructure is MARVEL’s main transfer vehicle, with Materials Cloud, AiiDA/AiiDAlab, and Quantum Mobile as the main success, as well as more than 40 open-source codes. These tools make complex simulations easier to run, easier to reproduce, easier to share and easier to teach. They help researchers keep track of how a result was produced, compare results from different simulation engines, publish reusable datasets, and turn expert workflow into tools that many more people can use.

The Materials Cloud

The Materials Cloud is a platform built to enable the seamless sharing and dissemination of resources in computational materials science, offering educational, research, and archiving tools, simulation software and services, and curated and raw data. The Materials Cloud have become the backbone for numerous European computing initiatives, ensuring its sustained development and maintenance through external consortia like BIG-MAP and SwissTwins. Materials Cloud Archive is now a fully established, long-term open-access repository officially recommended by the SNSF, Nature's Scientific Data, and Open Research Europe. The data retention has already been prepaid to guarantee online preservation for at least 10 years after submission. To support the increasing number of submissions and to enable long-term sustainability, MARVEL has transitioned its backend from a custom-built solution to the robust InvenioRDM framework (developed at CERN and the backbone of many large repositories including Zenodo). The maintenance of moderation is planned to be absorbed by major external projects, including support by CECAM, while the continuation of the technical maintenance of the Archive is being discussed with HPC centers including CSCS and CINECA (Italy).

AiiDA and AiiDAlab

AiiDA is an open-source Python infrastructure to help researchers with automating, managing, persisting, sharing and reproducing the complex workflows associated with modern computational science and all associated data. Many new AiiDA plugins and workflows have been developed, with 106 plugin packages supporting 180 simulation codes and more than 200 workflows, as of spring 2026. More than half of these were developed outside of MARVEL, showing the adoption by the community and demonstrating sustainability of AiiDA. AiiDA tutorials are regularly organized. Upcoming events can be found here.

AiiDAlab is a web environment for AiiDA allowing to run complex and robust AiiDA workflows with tailored lightweight web applications directly in the browser. It offers tailored applications for experimentalists and common standards such as OPTIMADE and reusable workflow interfaces are made available. A particularly convincing translational example came from Empa, where tailored AiiDAlab applications let experimentalists browse, launch and analyze workflows themselves, shifting the interaction with modelers away from technical setup and toward scientific interpretation. Now AiiDAlab has become a core and supported capability for different laboratories.

Quantum Mobile, OSSCAR and Lhumos, as open-source educational infrastructure

Quantum Mobile, a uniform software environment for teaching, with interactive Jupyter notebooks and virtual machines, has become a model for knowledge transfer in real and virtual training events.  Standard, but modular and customizable, these virtual machines have been used in many tutorials and events, with very good testimonials. 

The digital platform OSSCAR (Open Software Services for Classrooms and Research) is an educational resource developed with EPFL and CECAM. As Quantum Mobile, it will continue to serve global classrooms, with notebooks used in Bachelor and Master courses.  

Tools such as Quantum Mobile and OSSCAR lower the entry barrier for computational materials teaching.

The Materials Cloud Learn and Lhumos platforms act as repository of the recording of many activities turning them into lasting online resources, aiming to leave a lasting impact beyond the end of the NCCR. We can find there recording of many educational activities, such as the Distinguished Lectures, the Mary Ann Mansigh Conversations and the Classics in Molecular and Materials Modeling, and also many schools and tutorials.

All these are providing open-source educational infrastructure that is globally accessible and that will ensure that the educational legacy of MARVEL continues to train the next generation of computational materials scientists.

Open-source codes

MARVEL is also more than 40 open-source materials simulation codes, the major is listed here on the website. These are also at the center of trainings of users in advanced computational methods as, e.g., for Metatensor or Quantum ESPRESSO.