Execution Modes
After building Data Designer configurations with the library’s data_designer.config APIs, you choose where the NeMo Data Designer plugin executes the workload.
The important distinction is not simply “local” versus “remote”. There are two separate questions:
run versus submit primarily controls where the plugin workload execution happens. It does not necessarily determine whether the workload uses NeMo Services APIs.
nemo data-designer ... run can be fully local, but it is not an offline-only mode. A local run can still use the Files API, Secrets API, and Inference Gateway API from a running NeMo Services cluster when the configuration references the corresponding resources.
Terms
Execution Surfaces
Resource Compatibility
Local CLI execution supports the broadest set of resource patterns because the workload runs in your Python environment.
Choosing a Mode
Use CLI run when you want fast local iteration, direct access to local files, or library-equivalent workload behavior.
Use CLI submit when you want service-managed execution, job lifecycle tracking, shared artifacts, or the same execution environment other users and automation will use.
Use the SDK when you are writing application code that should call the Data Designer API programmatically. Today, the SDK path exclusively uses NeMo Services; use CLI run for local in-process execution.
If you run NeMo Services fully locally, submit and SDK calls can target that local deployment. The services are local, but the workload execution path still uses the Data Designer API and Jobs API.
Local NeMo Services Execution
To use submit or the SDK against a local NeMo Services deployment, run the services locally:
Run this from an environment with the service components and the Data Designer plugin installed. In another terminal, run nemo setup or nemo setup --auto to register your provider, then use preview submit, create submit, or the SDK against http://localhost:8080.