SimPilot
DocumentationExamplesAPI Reference
Sign in

Search Documentation

Find pages across the documentation

Getting Started

IntroductionQuickstartHow It Works

Core Concepts

AI AgentSimulation WorkflowEngineering PipelineKnowledge BaseArchitecture

Features

OverviewChat InterfaceCanvas & ArtifactsSharing & CollaborationExports & ReportsVoice InputMulti-Model AIProjectsTemplates & WorkflowsFile UploadsDeep ResearchWeb SearchCode ExecutionImage GenerationMCP ConnectorsInteractive DashboardsParallel ExecutionURL RetrievalFreeCAD CADSimulationsTemplatesConvergence Monitoring

Simulation

OverviewSupported TypesMesh GenerationError RecoveryBatch & SweepsResults Comparison

Studies & Analysis

OverviewDOE & Parametric SweepsOptimizationComparison

Validation & Reviews

OverviewBaselines & VersioningEngineering ReviewsRegulatory Compliance

3D Viewer

OverviewVisualization ToolsKeyboard Shortcuts

Enterprise

OverviewAdmin PanelOrganizationsKnowledge ManagementMethod Packs

Account

Getting StartedSettingsBilling

Examples

OverviewAerodynamicsPipe FlowHeat TransferStructural
  1. Docs
  2. Simulation
  3. Simulation Overview

Simulation Overview

SimPilot's simulation capabilities -- from CFD to structural analysis.

SimPilot provides a full-stack simulation environment built on open-source solvers and cloud infrastructure. You describe a problem in natural language, and the AI agent selects the right solver, generates the mesh, creates all configuration files, executes the run, and delivers results -- automatically.

Plugin-based software system

SimPilot uses a plugin architecture for simulation software. Each software plugin implements a standardized interface that covers case setup, mesh generation, file validation, error detection, result extraction, and post-processing. This design makes it straightforward to add new software without modifying the core orchestration engine.
Every plugin declares its own:
  • Allowed commands and command validation
  • Case directory structure and file generation order
  • Command evaluation, solver-aware failure detection, and retry-aware recovery guidance
  • Mesh quality thresholds and quality parsing via checkMesh tool
  • Result extraction logic
  • Knowledge databases for RAG-guided troubleshooting

Registered software

SimPilot currently ships with 3 registered software plugins:
SolverStatusDomain
OpenFOAMFull supportCFD, heat transfer, multiphase, reactive flow
SU2Full supportCompressible aerodynamics, adjoint optimization
CalculiX (CCX)Full supportStructural FEA and thermal stress analysis
All software plugins production-ready
All three software plugins are production-ready with complete case setup, mesh pipeline, error recovery with evaluator-optimizer diagnosis and retry history, result extraction, and post-processing support.

Runtime backends

Simulations can execute on 4 runtime backends, selected via configuration:
laptop

Local

Runs solver commands directly on the host machine. Used for development and testing.
box

Docker

Executes inside a Docker container with the solver pre-installed. Default image includes OpenFOAM v2512.
server

HPC

Submits jobs to SLURM and PBS job schedulers via SSH, with resource templates for common cluster configurations. Requires HPC host configuration.
cloud

Remote (AWS ECS)

Cloud execution on managed containers via the SimPilot Compute API. This is the production default for all hosted users.

Cloud execution

In production, simulations run on Docker containers deployed on AWS ECS. The compute image ships with OpenFOAM v2512 (ESI), GMSH, Python 3, CadQuery, PyVista, and all mesh utilities pre-installed. The remote runtime handles file transfer, job submission, progress polling, and result retrieval transparently.
The AI agent streams real-time progress updates -- including solver stage detection (meshing, solving, post-processing), Courant number tracking, and residual convergence -- back to the chat interface as the simulation runs.

Real-time monitoring

During simulation runs, SimPilot monitors convergence in real time -- tracking residuals, Courant numbers, and solver progress. It detects divergence, stagnation, and oscillation patterns, and provides coaching suggestions (e.g., "Residuals stalling -- consider reducing relaxation factors"). Health snapshots with ETA are streamed to the chat.

Post-processing

The post-processing phase of the engineering pipeline provides standalone post-processing for completed simulations. It generates 3D visualizations (contour plots, streamlines, slices) via PyVista and 2D plots (residuals, convergence) via matplotlib, with results rendered as inline images in the chat and available for download in PNG, SVG, and PDF formats. For interactive data dashboards, see Interactive Dashboards.

Explore the simulation system

flask

Supported Simulation Types

18 OpenFOAM solvers across 7 physics categories, plus SU2 and CalculiX (CCX) plugin coverage.
grid-2

Mesh Generation

Four mesh generation modes from automatic blockMesh to agentic GMSH scripting and geometry pipelines.
wrench

Error Recovery

Evaluator-optimizer recovery, retry history, optional error-diagnostician delegation, and targeted web-search escalation.
chart-bar

Batch & Parameter Sweeps

Run up to 20 simulations in parallel with automated comparison and parameter sweeps.
scale

Results Comparison

Side-by-side metric tables, regression alerts, and best/worst rankings across batch runs.
PreviousConvergence MonitoringNextSupported Types

On this page

Plugin-based software systemRegistered softwareRuntime backendsCloud executionReal-time monitoringPost-processingExplore the simulation system