Gaussian Backend
Gaussian is a quantum-chemistry backend available through PyAR for users who have a licensed Gaussian installation and want higher-level electronic- structure calculations.
When to use Gaussian
Use gaussian when:
Gaussian is installed and accessible in your environment
you want DFT-style optimisation or refinement
you are refining a smaller set of candidates generated by cheaper PyAR searches
you have validated a small test calculation with your method, basis, charge, and multiplicity
Example command
Reaction search using the registered Gaussian energy-gradient route:
pyar-cli react A.xyz B.xyz -N 8 -gmin 100 -gmax 1000 --software gaussian --method B3LYP --basis 6-31G*
Practical notes
Gaussian must be installed separately and configured according to your local licence and system setup.
Confirm the executable name and output handling on a small molecule before running larger PyAR jobs.
Gaussian calculations are usually much slower than xTB or ML-potential runs. For screening, use xTB/AIMNet2 first and refine selected structures later.
AFIR/geomeTRIC status
gaussian is wired as a registered energy-gradient provider for the
geomeTRIC-backed AFIR reaction route. This means PyAR can construct a combined
Gaussian energy/gradient + AFIR bias objective for geomeTRIC. Local
validation is still required because Gaussian setup and output handling depend
on the installed executable and environment.