FlexSim MCP Server
Control and automate FlexSim simulations through AI assistants like Claude, enabling manufacturing and warehouse digital twin analysis, parameter studies, and real-time model manipulation via tools for opening models, running simulations, evaluating FlexScript, and exporting results.
README
title: FlexSim MCP Server emoji: ⚙️ colorFrom: blue colorTo: green sdk: gradio sdk_version: "5.0.0" app_file: app.py pinned: false
FlexSim MCP Server (incl. FlexSimPy)
Control and automate FlexSim simulations via the Model Context Protocol (MCP). This repository exposes a FastMCP server that launches FlexSim (via FlexSimPy) and provides tools to open models, run/stop, step, query/set node values, evaluate FlexScript, export results, and more.
Why this matters
- Bridges digital twins and AI assistants (Claude/Cursor) for manufacturing/warehouse analysis and “what-if” studies.
- Demonstrates robust systems engineering: GUI process orchestration, async I/O, protocol handling, Windows quirks, and reproducible automation.
Design
┌─────────────────────┐ JSON-RPC ┌─────────────────────┐
│ MCP Client │◄────────────────────► │ MCP Server │
│ (Claude/Cursor) │ stdin/stdout │ (flexsim_mcp.py) │
└─────────────────────┘ └──────────┬──────────┘
│
│ Python API
▼
┌─────────────────────┐
│ FlexSimPy │
│ (Python SDK) │
└──────────┬──────────┘
│
│ COM/IPC
▼
┌─────────────────────┐
│ FlexSim │
│ (Simulation Engine)│
└─────────────────────┘
Features
- MCP server entry point:
mcp_server/flexsim_mcp.py - Tools exposed:
- Simulation control:
flexsim_open_model,flexsim_reset,flexsim_run,flexsim_run_to_time,flexsim_stop,flexsim_step,flexsim_get_time - Model/script:
flexsim_evaluate,flexsim_compile,flexsim_save_model,flexsim_new_model - Node access:
flexsim_get_node_value,flexsim_set_node_value - Results/stats:
flexsim_get_statistics,flexsim_export_results
- Simulation control:
- Logging:
mcp_server/flexsim_mcp.log - Reproducible environment:
uv sync - FlexSimPy build automation:
utility/build_automation.py - Integration tests:
tests/integration/test_mcp_client.py(stdio-based MCP client) - GUI sanity test:
utility/integration_test.py - Configurable behavior via
config.tomland environment overrides
Requirements
- OS: Windows
- Python: 3.12 (default; configurable via
config.tomlorFLEXSIM_PYTHON_VERSION) - Package manager:
uv(https://astral.sh) - FlexSim:
- Real runs require a local FlexSim installation (or repo-local mirror in
FlexSimDev/program) - FlexSimPy SDK submodule is provided at
depends/FlexSimPy(see.gitmodules)
- Real runs require a local FlexSim installation (or repo-local mirror in
Quickstart
Guided setup (recommended)
# From the repo root
python setup_mcp_server.py
# The setup helper will:
# - Summarize your configuration (install locations, Python version, submodule presence)
# - Ensure `uv` is installed and run `uv sync`
# - Build/install FlexSimPy via `utility/build_automation.py`
# - Optionally run a GUI integration sanity test
# - Print next steps to run the MCP server and tests
Manual setup
# 1) Install dependencies
uv sync
# 2) Check (and optionally build) FlexSimPy
uv run python utility/build_automation.py --status
# To force a build with a specific Python version:
# uv run python utility/build_automation.py --python-version 3.12
# 3) Launch MCP server in test mode (keeps the process alive; launches FlexSim on first tool call)
uv run mcp_server/flexsim_mcp.py --test-mode
# 4) Run integration tests (spawns the server as a subprocess; stdio MCP client)
uv run python tests/integration/test_mcp_client.py
Configuration
Primary configuration lives in config.toml (repo root). You can also override at runtime using environment variables.
Example (excerpt from config.toml):
[flexsim]
install_path = "FlexSimDev/program"
src_path = "C:\\Program Files\\FlexSim 2025 Update 2"
[python]
version = "3.12"
[server]
name = "FlexSimPy MCP Server"
version = "0.1.0"
http_endpoint = "http://127.0.0.1:8088/mcp"
[session]
reuse_policy = "singleton"
[logging]
level = "INFO"
log_file = "flexsim_mcp.log"
[build]
flexsimpy_dir = "depends/FlexSimPy"
auto_build = true
auto_install = true
Environment overrides (see utility/config.py):
FLEXSIM_CONFIG_PATH– path to an alternateconfig.tomlFLEXSIM_INSTALL_PATH– override FlexSim program pathFLEXSIM_PYTHON_VERSION– override Python version for FlexSimPyFLEXSIM_LOG_LEVEL– override logging level
Running the MCP server
Two modes:
- Test mode (interactive; no stdio protocol)
uv run mcp_server/flexsim_mcp.py --test-mode
# Keeps running; FlexSim launches on first tool call; intended for manual testing
- MCP stdio mode (default when no flags are passed)
uv run mcp_server/flexsim_mcp.py
# Expects MCP JSON-RPC messages via stdin/stdout; used by clients/tests
Client configuration (for Claude/Cursor/etc.):
{
"mcpServers": {
"flexsim": {
"command": "uv",
"args": [
"--directory",
"<repo-flexsim_mcp>/",
"run",
"mcp_server/flexsim_mcp.py"
]
}
}
}
Example tool calls (JSON-RPC via MCP)
Open a model
{"jsonrpc":"2.0","id":1,"method":"tools/call",
"params":{
"name":"flexsim_open_model",
"arguments":{"params":{
"model_path":"C:/path/to/Model.fsm"
}}
}}
Run to time and read time
{"jsonrpc":"2.0","id":2,"method":"tools/call",
"params":{"name":"flexsim_run_to_time","arguments":{"params":{"target_time":300}}}}
{"jsonrpc":"2.0","id":3,"method":"tools/call",
"params":{"name":"flexsim_get_time"}}
Evaluate FlexScript
{"jsonrpc":"2.0","id":4,"method":"tools/call",
"params":{"name":"flexsim_evaluate","arguments":{"params":{
"script":"getmodeltime()"
}}}}
Get/Set node value
{"jsonrpc":"2.0","id":5,"method":"tools/call",
"params":{"name":"flexsim_get_node_value","arguments":{"params":{
"node_path":"Model/Queue1/stats/input"
}}}}
{"jsonrpc":"2.0","id":6,"method":"tools/call",
"params":{"name":"flexsim_set_node_value","arguments":{"params":{
"node_path":"Model/Processor1/variables/processtime",
"value":5.0
}}}}
More FlexScript examples and behavior: see eval-flexscript.md.
Tests
Integration tests (stdio client; verifies tools and basic run sequence):
uv run python tests/integration/test_mcp_client.py
# Test 1: open model; Test 2: run for 20s, stop, verify time > 0
GUI sanity test (manual, interactive menu):
uv run python utility/integration_test.py
# Launches FlexSim GUI and provides simple interactive actions
Repository structure
mcp_server/flexsim_mcp.py— FastMCP server and FlexSimPy controller logic (entry point)utility/config.py— configuration loader with env overridesbuild_automation.py— FlexSimPy build/install orchestration (MSBuild resolution, .pyd install)copy_flexsim.py— copies a FlexSim installation intoFlexSimDev/(usesflexsim.src_path)integration_test.py— GUI-backed sanity test (suppresses CEF stderr, safe cleanup)utility.py— helpers: suppress stderr, kill FlexSim processes, force exit
tests/integration/test_mcp_client.py— async stdio client that starts the server and calls toolsREADME.md— test descriptions and sample outputs
depends/FlexSimPy— FlexSimPy SDK submoduleFlexSimDev/program— optional repo-local mirror of FlexSim installconfig.toml,mcp_server_config.json,AGENTS.md,eval-flexscript.md
Batch Experiment Orchestrator (Design)
A high-impact enhancement to run parameter sweeps with replications, gather metrics, and export a unified summary.csv (and manifest.json), enabling “what-if” studies at scale.
- Design document:
Batch_Experiment_Orchestrator.md - Status: design in-repo; implementation planned (not yet in
flexsim_mcp.py) - MVP tool (planned):
flexsim_run_experimentswith parameter list, seeds/replications, run_to_time, stats, artifact_dir
Troubleshooting
- FlexSim not found:
- Update
flexsim.install_pathinconfig.tomlor setFLEXSIM_INSTALL_PATH - Optionally use
utility/copy_flexsim.pyto mirror an existing install intoFlexSimDev/
- Update
- FlexSimPy unavailable:
- Check status:
uv run python utility/build_automation.py --status - Rebuild for the active Python version and ensure the
.pydis installed
- Check status:
- Stdio client issues:
- Ensure you run the server without
--test-modewhen using stdio clients - Check
mcp_server/flexsim_mcp.logfor details
- Ensure you run the server without
- GUI noise (CEF) or shutdown issues:
- The utilities suppress CEF stderr and force a safe interpreter exit when needed
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.