aflow

AFLOW, short for Automatic FLOW, is a high-throughput computational framework and accompanying database designed to accelerate the discovery and design of crystalline materials. It automates and standardizes the generation of crystal structures, the execution of first-principles electronic-structure calculations, and the curation of resulting data to enable large-scale screening and data-driven analysis.

Core components include an automated workflow engine that handles structure prototypes, symmetry analysis, and convergence testing; a standardized set of calculation parameters and metadata to ensure comparability across materials; and routines for extracting properties such as formation energy, lattice parameters, band gaps, electronic density of states, elastic constants, and, where available, vibrational and thermodynamic data. AFLOW also maintains a prototype library for known crystal structures and a structured data model to describe materials in a uniform way.

AFLOWlib is the public data repository derived from AFLOW workflows, providing programmatic access via an API and the AFLUX search language for complex queries. The framework originated from the AFLOW consortium, notably developed at Duke University, by researchers including Stefano Curtarolo and collaborators, and has become widely used in the materials science community for high-throughput screening, data-driven materials discovery, and machine learning model development.