Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@su2-mcp Run a subsonic Euler simulation on my wing mesh"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
su2-mcp exposes SU2 session lifecycle, config editing, solver execution, and
results inspection through an MCP server.
Session lifecycle management with per-session working directories.
Config helpers for reading, parsing, and updating SU2 .cfg files.
Mesh upload and automatic MESH_FILENAME synchronization.
Optional STEP -> SU2 mesh conversion via gmsh (generate_mesh_from_step).
Solver wrappers for SU2_CFD and SU2_DEF with timeout/missing-binary
handling.
Result-file listing, base64 download, history CSV parsing, and surface sampling.
Open-ended mesh refinement (2026-06-21): the CPACS adapter exposes
optional surface_density and farfield_factor overrides on top of
the three named presets (laptop/workstation/industry), and
reports the actual mesh_n_elem cell count plus a cauchy_triggered
flag so callers can build a converge-until-plateau loop (see the
SKILL_OPEN_ENDED_MESH.md skill and scripts/run_converged_su2.py
in cmudrc/agent-mcp).
Related MCP server: tigl-mcp
python3 -m venv .venv
source .venv/bin/activate
pip install -e .[dev]su2-mcp --transport stdiosu2-mcp --transport http --host 0.0.0.0 --port 8002 --path /mcp
su2-mcp --transport streamable-http --host 0.0.0.0 --port 8002 --streamable-http-path /mcpsu2-mcp --transport sse --host 0.0.0.0 --port 8000 --mount-path / --sse-path /sseProgrammatic entrypoints are available from su2_mcp.main.
make dev
make lint
make type
make test
make docsFull local CI-style gate:
make ciThe MCP tool get_su2_status reports availability of common SU2 binaries
(SU2_CFD, SU2_CFD_MPI, SU2_DEF).
pip install .[su2]The Python package is pure Python, but some tools rely on external executables:
run_su2_solver / generate_deformed_mesh: requires SU2 binaries on PATH
generate_mesh_from_step: requires gmsh on PATH
This MCP includes a CPACS adapter (src/su2_mcp/cpacs_adapter.py) that
bridges SU2 to the shared-CPACS aircraft analysis pipeline.
The adapter reads reference geometry and flight conditions from CPACS, meshes
STEP files via Gmsh, runs real SU2_CFD Euler simulations, parses CL/CD from
history.csv, and writes aerodynamic results into //analysisResults/aero.
Direction | XPath |
Reads |
|
Writes |
|
Records | one |
# Classic preset path
python pipeline/shared_cpacs_orchestrator.py D150_v30.xml \
--mcps tigl su2 pycycle nseg --su2-preset workstation
# Custom one-off density (open-ended override)
python pipeline/shared_cpacs_orchestrator.py D150_v30.xml \
--mcps tigl su2 --su2-density 120
# Converged delivery: refine surface_density until CL/CD plateau
python pipeline/shared_cpacs_orchestrator.py D150_v30.xml \
--mcps tigl su2 --su2-convergeSee cmudrc/aircraft-analysis for full pipeline documentation, versioning details, and installation instructions.
MCP | Repository |
TiGL (geometry) | |
pyCycle (engine cycle) | |
Mission, segment/Breguet | |
Mission, trajectory-level |
Mayank Dixit (@Kugel-Blitz-13), Carnegie Mellon University — mayankd@cmu.edu Christopher McComb, Carnegie Mellon University — Design Research Collective
This server cannot be deployed
Design, solve and simulate HVAC systems from real components, weather years and buildings.
Real BEM engineering for propellers, wings and turbines — design, optimize, visualize, export.
Control Unreal Engine to browse assets, import content, and manage levels and sequences. Automate…
Governed data discovery, exact queries, decisions, simulations, and runtime utilities over MCP.
