2. How to use Wake CFD

Wake CFD provides a desktop GUI, command-line interface (CLI), and Python API for case authoring and execution. All interfaces share the same JSONC input model, so a case prepared in one interface can be reused in another execution environment.

Each interface is designed for different stages and purposes of an analysis workflow rather than being interchangeable tools.

Wake CFD interface workflow from interactive authoring through HPC execution to automated sweeps

Placeholder for the user-interface workflow. Before public release, replace with a diagram showing GUI → CLI → Python API paths and their typical use cases.

Choosing an interface

Tip

First-time users

If you have not yet run Wake CFD, complete the packaged example in Getting started to verify that Wake CFD executes correctly.

Interface Best suited for Notes
GUI Mesh-marker inspection, interactive case authoring, and setting exploration Visual case setup is available. Post-processing visualization is under development; currently use external tools such as ParaView.
CLI Reproducible runs, parameter sweeps, HPC/MPI, and version control You manage JSONC files, execution commands, and file paths directly. Post-process results in ParaView.
Python API Case generation, repeated runs, parameter sweeps, and result-processing automation Requires a Python environment and Wake CFD SDK setup.

Common workflows

Interactive case authoring Use Wake CFD Desktop to inspect mesh zones and boundary markers, then configure physics models, boundary conditions, numerical methods, and outputs. Run locally or export the case to JSONC.

GUI to HPC transition Author or review a case in Wake CFD Desktop, then export it to JSONC. Run the same case on a remote GPU server, MPI environment, or Slurm cluster using the CLI.

Repeated analysis and parameter sweeps Start from a packaged example or existing JSONC case. Use CLI overlays, patches, or the Python API to vary Mach number, angle of attack, mesh, and numerical settings across repeated runs.

External-system integration Submit cluster jobs with Slurm, monitor or control running solvers through gRPC, or use an MCP-based agent to author cases and attach to the live Desktop project. The agent and GUI share the same project model, so automated edits can be reviewed in the open interface. See Automation and integration.

Settings shared across all interfaces

The following items use the same case model in the GUI, CLI, and automation environments:

  • Mesh files and zone/boundary markers
  • Analysis method and physical models
  • Materials and reference values
  • Domains and boundary conditions
  • Domain interfaces
  • Numerical methods and run control
  • Initial conditions
  • Monitoring and file outputs

For the physical meaning and JSONC keys of each setting, see Input and configuration reference. The GUI edits these settings visually; the CLI and automation interfaces use the same content in JSONC.

Next steps