Usage examples#

These examples show typical MCP workflows once the server is configured.

Load a case and summarize setup#

User prompt:

“Connect to Fluent, load ``pipe.cas.h5``, and summarize the current setup.”

Expected tool sequence:

  1. connect launches or attaches to Fluent.

  2. run_code calls session.file.read_case(file_name="pipe.cas.h5").

  3. summarize_setup returns a compact digest of models, boundary conditions, materials, and numerics.

Inspect mesh quality#

User prompt:

“Show the mesh quality for this case, including skewness and orthogonal quality.”

Expected tool:

mesh_quality(include_check=False)

Use mesh_quality(include_check=True) when you also need Fluent’s mesh.check() output.

Generate and apply a boundary condition change#

User prompt:

“Set the velocity inlet ``inlet-1`` to 15 m/s.”

Expected tool sequence:

  1. list_named_objects("setup.boundary_conditions.velocity_inlet") confirms the name.

  2. get_state("setup.boundary_conditions.velocity_inlet") reads current inlet settings.

  3. Generate a PyFluent settings snippet in the external orchestration layer.

  4. validate_code pre-checks the snippet.

  5. run_code applies the deterministic settings API calls.

  6. get_state("setup.boundary_conditions.velocity_inlet") verifies the new value.

Perform offline API discovery#

User prompt:

“What settings path controls the turbulence model when Fluent is not running?”

Expected tool sequence:

  1. find_api("turbulence model") performs ranked path search offline.

  2. get_help("setup.general.turbulence.model") returns per-path help text.

Compare two case files#

User prompt:

“Compare ``baseline.cas.h5`` and ``modified.cas.h5`` and summarize what changed.”

Expected tool:

compare_files(path_a="baseline.cas.h5", path_b="modified.cas.h5")

Reply with the summary markdown table from the response verbatim.

Next steps#