“…During the last decade, the interest in workflow development has grown considerably in the scientific community 66 and multi-purpose engines for managing calculation workflows, have been developed, including AFLOW 23,67,68 , AiiDA 31,69 , ASE 34 , and Fireworks 70 . Using these infrastructures, a number of workflows have been used for scientific purposes, like convergence studies 71 , equations of state (e.g., AFLOW Automatic Gibbs Library 72 and the AiiDA common workflows ACWF 73 ) , phonons [74][75][76][77] , elastic properties (e.g., the elastic-properties library for Inorganic Crystalline Compounds of the Materials Project 78 , AFLOW Automatic Elasticity Library, AEL 79 , ElaStic 80 ), anharmonic properties (e.g., the Anharmonic Phonon Library, APL 81 , AFLOW Automatic Anharmonic Phonon Library, AAPL 82 ), highthroughput in the compositional space (e.g., AFLOW Partial Occupation, POCC 83 ), charge transport (e.g., organic semiconductors 84,85 ), of covalent organic frameworks (COFs) for gas storage applications 86 , of spindynamics simulations 87 , high-throughput automated extraction of tight-binding Hamiltionians via Wannier functions 88 , and high-throughput on-surface chemistry 89 There are two types of metadata associated to workflows. Thinking of a workflow as a code to be run, the first type of metadata characterizes the code itself.…”