Environment Variables

Configure NIM for BMD behavior at container startup using the following environment variables.

Required Variables

Variable	Required	Default	Description
`NGC_API_KEY`	Yes (for MACE auto-download)	—	NGC authentication key. Required for the container to download the bundled MACE model from NGC.

Configuration Variables

Variable	Type	Default	Description
`ALCHEMI_NIM_MODEL_TYPE`	string	`"mace"`	Model type: `mace`, `aimnet2`, or `tensornet`.
`ALCHEMI_NIM_MODEL_PATH`	string	null	Path to the model file or directory inside the container.
`ALCHEMI_NIM_PBC`	bool	true	Periodic boundary conditions. If true, requests must include `cell`. If false, requests must not include `cell`.
`ALCHEMI_NIM_BATCH_SIZE`	int	0	Batch size in atoms (0 enables auto-estimation at startup).
`ALCHEMI_NIM_BATCH_SIZE_ESTIMATE_DATA`	string	`"omat24"`	Dataset used for auto-estimation: `omat24` (bulk materials), `omol25` (molecules), or `omc25` (molecular crystals).
`ALCHEMI_NIM_DFT3_ENABLED`	bool	false	Enable DFT-D3(BJ) dispersion correction.
`ALCHEMI_NIM_DFT3_PARAM`	string (JSON)	null	JSON string containing custom DFT-D3(BJ) parameters (for example, `{"s8": 0.3908, "a1": 0.5660, "a2": 3.1280}`). Refer to the ALCHEMI Toolkit-Ops documentation for details.

Container Runtime Variables

Variable	Required	Default	Description
`NIM_HTTP_API_PORT`	No	`8000`	HTTP port for the inference API.
`NIM_DISABLE_MODEL_DOWNLOAD`	No	false	Whether to automatically download the default MACE model from NGC.

For launch examples, refer to Custom Models and Quickstart Guide.

Configuration for configuration options, per-request overrides, and model configuration
Supported Models for model sourcing