ChemistryCPU4 credits

MACE-MP-0 — Universal Interatomic Potential

DFT-quality molecular dynamics for 89 elements

MACE-MP-0 is a universal neural network potential covering 89 elements, trained on the Materials Project. Replaces DFT for molecular dynamics at 1000× the speed. Developed by Cambridge ACEsuit, MIT licensed.

$0.04
per API call
4
credits per call
/v1/chemistry/qm/properties
API endpoint

Features

89-element coverage
Equivariant message-passing architecture
DFT-quality energies and forces
Battery and catalyst screening
CPU-friendly for small systems

Quick Start

MACE-MP-0 — Python Examplepython
import requests

result = requests.post(
    "https://scirouter.ai/v1/chemistry/qm/properties",
    json={"smiles": "CC(=O)OC1=CC=CC=C1C(=O)O", "method": "mace-mp-0"},
    headers={"Authorization": "Bearer sk-sci-your-key"}
).json()["data"]

Use Cases

1

Battery electrode screening

2

Catalyst optimization

3

Surface reaction MD

4

Crystal structure prediction

Start Using MACE-MP-0

500 free credits every month. No credit card required.