MACE, CHGNet, Nequix, SevenNet, Orb, MatterSim universal machine learning interatomic potentials (u-MLIPs)

Use MACE, CHGNet, SevenNet, Nequix, Orb, MatterSim, or PET-MAD universal machine learning interatomic potentials (uMLIP) in interactive interface for atomistic simulations

If you like the app, please cite: Lebeda, M., Drahokoupil, J., Mazáčová, V., & Vlčák, P. (2026). Revealing interstitial energetics in Ti-23Nb-0.7Ta-2Zr gum metal base alloy via universal machine learning interatomic potentials. Journal of Materials Research and Technology, 41, 6766–6774.

Compile the application from GitHub: https://github.com/bracerino/uMLIP-Interactive

Universal MACE, CHGNet, SevenNet, Nequix, Orb, MatterSim, and PET-MAD machine learning interatomic potentials for calculations of single-point energies, geometry optimization, phonons, elastic properties, genetic algorithm for substitutes / vacancies, and molecular dynamics using an in interactive interface.

Multiple structures can be uploaded at once and the calculations run in batch mode.

For online version which can be used to generate standalone python scripts for selected calculations, please visit: https://umlip-interactive.streamlit.app