Command Palette

Search for a command to run...

UnylyUnyly
Весь каталог

Openems

БесплатноНе проверен

Enables AI-assisted design of antennas and electromagnetic structures using openEMS FDTD simulation, with analytical dimension calculators and full-wave script

GitHubEmbed

Описание

Enables AI-assisted design of antennas and electromagnetic structures using openEMS FDTD simulation, with analytical dimension calculators and full-wave script generation.

README

mcp-openems: 3D FDTD electromagnetic simulation

License Python 3.10+ MCP eng-mcp-suite

AI-assisted antenna and RF-structure design via openEMS FDTD, driven over MCP. Patch / dipole / monopole / horn / helix antennas, microstrip and coupled-line transmission lines, via transitions: geometry, analytical Z₀/εeff, and ready-to-run openEMS Python scripts.

Quick start · Tools · Examples · Suite catalog


An MCP server for designing antennas and electromagnetic structures using openEMS FDTD simulation. Provides analytical design calculators that work immediately, plus openEMS script generation for full-wave simulation.

Part of the engineering toolkit

This repo is part of eng-mcp-suite, an MCP-driven engineering toolkit for RF / EMC / PCB / signal-integrity / lab-test workflows.

Related tools in the toolkit:

Tool When to reach for it
lineforge 2D quasi-TEM closed-form for transmission lines (microstrip, stripline, CPWG, differential, three-conductor). Use this when you need impedance fast and the geometry is 2D.
mcp-nec2-antenna Wire-antenna method-of-moments (dipole / Yagi / vertical / loop / inverted-V). Use this when you have a wire-antenna geometry and don't need the 3D-field detail of FDTD.
mcp-pcb-emcopilot PCB layout review (decoupling, return paths, plane resonances, DDR/PCIe/USB SI). Often pairs with mcp-openems for full-wave validation of a flagged region.

When to use mcp-openems specifically: full-wave 3D FDTD validation, broadband S-parameters, near/far-field characterization, antenna geometries with 3D features (horns, helices), or when closed-form is running out of accuracy.

Features

Antenna Design Tools

  • openems_create_patch - Microstrip patch antenna (WiFi, GPS, satellite)
  • openems_create_dipole - Half-wave dipole antenna
  • openems_create_monopole - Quarter-wave monopole over ground plane
  • openems_create_horn - Pyramidal horn antenna for specified gain
  • openems_create_helix - Axial-mode helix for circular polarization

Simulation & Export

  • openems_generate_script - Generate complete OpenEMS Python simulation script
  • openems_check_installation - Check if OpenEMS is available

Design Management

  • openems_list_designs - List all designs in session
  • openems_get_design - Get full design details with geometry
  • openems_list_antenna_types - Reference for antenna types and applications

Installation

1. Clone and install

git clone https://github.com/RFingAdam/mcp-openems.git
cd mcp-openems
uv pip install -e .

2. (Optional) Install OpenEMS for simulation

The MCP works without OpenEMS: the design tools calculate dimensions analytically. Full FDTD simulation needs the CSXCAD and openEMS Python bindings.

These are not on PyPI. pip install CSXCAD openEMS does not work. They are compiled extensions built from source against the openEMS C++ install:

git clone --recursive https://github.com/thliebig/openEMS-Project.git
cd openEMS-Project
./update_openEMS.sh ~/opt/openEMS          # builds the C++ libraries
source .venv/bin/activate                  # the venv you installed this MCP into
./scripts/build_python.sh --cpp-install-dir ~/opt/openEMS

Two things that will bite you, both of which broke this repo in 2026-08:

  • The bindings are ABI-pinned to one Python minor version. A cpython-314 extension is invisible to a 3.12 interpreter, and vice versa. If you recreate the venv on a different Python, rebuild the bindings too.
  • They link against system HDF5, VTK and boost. A distro upgrade that bumps those sonames breaks every .so here with libhdf5_serial.so.NNN: cannot open shared object file. Rebuild against the new libraries; build_python.sh --cpp-install-dir rebuilds only the Python layer, which is enough when the C++ install is already current.

Run openems_check_installation to see which of these you are hitting: it reports the real import error and the interpreter in use.

3. Add to your MCP client

Claude Code:

claude mcp add openems -- uv run --directory /path/to/mcp-openems mcp-openems

Config file format:

{
  "command": "uv",
  "args": ["run", "--directory", "/path/to/mcp-openems", "mcp-openems"]
}

Usage Examples

Design a WiFi patch antenna

Design a 2.4 GHz patch antenna on FR-4 substrate (er=4.4, 1.6mm thick)

The AI will:

  1. Use openems_create_patch to calculate dimensions
  2. Return patch length, width, feed position
  3. Provide estimated directivity and impedance

Design a satellite uplink antenna

I need a circularly polarized antenna for 5.8 GHz with about 12 dBi gain

The AI will use openems_create_helix for CP requirements.

Generate simulation script

Generate an OpenEMS script for this antenna so I can run a full simulation

The AI will use openems_generate_script to create a complete Python script.

Compare antenna types

What antenna types are available? I need something for a handheld radio at 440 MHz

The AI will use openems_list_antenna_types and recommend appropriate options.

Tool Reference

Design Output Format

Each design tool returns:

  • design_id: UUID for referencing the design
  • dimensions: Calculated physical dimensions in mm
  • calculated: Derived parameters (impedance, gain estimates)
  • geometry: OpenEMS-compatible geometry specification

Example Output

{
  "success": true,
  "design_id": "550e8400-e29b-41d4-a716-446655440000",
  "design": {
    "name": "2.4 GHz Patch",
    "type": "patch",
    "frequency_ghz": 2.4,
    "dimensions": {
      "patch_length_mm": 28.85,
      "patch_width_mm": 37.24,
      "feed_inset_mm": 8.92,
      "ground_plane_mm": 94.48
    },
    "calculated": {
      "effective_er": 3.33,
      "estimated_directivity_dbi": 7.2
    }
  }
}

Generated Script

The openems_generate_script tool creates a complete Python script that:

  1. Sets up the FDTD simulation
  2. Creates geometry from the design
  3. Adds mesh with appropriate resolution
  4. Runs the simulation
  5. Extracts S-parameters and plots results

Antenna Design Formulas

Antenna Method Key Formula
Patch Transmission Line Model L = c/(2f√εeff) - 2ΔL
Dipole Classical L = 0.95 × λ/2
Monopole Image Theory H = 0.95 × λ/4
Horn Aperture Theory G = 4πAe/λ²
Helix Kraus Model C ≈ λ, S = C tan(α)

Without OpenEMS

Even without OpenEMS installed, this MCP provides:

  • Analytical dimension calculations
  • Geometry specifications for manual modeling
  • Reference impedance and gain estimates
  • OpenEMS script generation for later use

With OpenEMS

With OpenEMS installed, you can run the generated scripts to:

  • Perform full-wave FDTD simulation
  • Get accurate S-parameters and input impedance
  • Calculate radiation patterns and gain
  • Visualize fields in ParaView

Supported Frequencies

The design tools work across the RF spectrum:

  • HF (3-30 MHz): Dipole, monopole
  • VHF (30-300 MHz): All types
  • UHF (300 MHz-3 GHz): All types
  • Microwave (3-30 GHz): Patch, horn, helix
  • mmWave (30-300 GHz): Patch, horn (with appropriate substrate)

License

AGPL-3.0-or-later. Relicensed from Apache-2.0 in v0.2.0 to align with the eng-mcp-suite toolkit-wide AGPL move. The underlying openEMS engine remains GPL-3.0; this wrapper is AGPL-3.0-or-later and invokes the engine at runtime without redistribution.

Trademarks and brand assets

This project is not affiliated with, endorsed by, or sponsored by the openEMS project. "openEMS" is used here only to identify the software this project interoperates with.

The project name and the logo files in this repository are not part of the licensed work. The licence above grants no permission to use them, except as needed to describe the origin of the work.

Author

Adam Engelbrecht - @RFingAdam

from github.com/RFingAdam/mcp-openems

Установка Openems

У этого сервера нет опубликованного пакета — он собирается из исходников. Открой репозиторий и следуй инструкции в README.

▸ github.com/RFingAdam/mcp-openems

FAQ

Openems MCP бесплатный?

Да, Openems MCP бесплатный — установка в пару кликов через Unyly без оплаты.

Нужен ли API-ключ для Openems?

Нет, Openems работает без API-ключей и переменных окружения.

Openems — hosted или self-hosted?

Self-hosted: сервер запускается локально на твоей машине командой из раздела установки.

Как установить Openems в Claude Desktop, Claude Code или Cursor?

Открой Openems на unyly.org, выбери вкладку своего клиента (Claude Desktop, Claude Code, Cursor) и нажми Install — конфиг сгенерируется автоматически, без правки JSON.

Похожие MCP

LibreOffice Tools

Enables AI agents to read, write, and edit Office documents via LibreOffice with token-efficient design. Supports multiple formats including DOCX, XLSX, PPTX, a

passerbyflutterавтор: passerbyflutter

dannote/figma-use

Full Figma control: create shapes, text, components, set styles, auto-layout, variables, export. 80+ tools.

dannoteавтор: dannote

Logo.dev

Search and retrieve company logos by brand or domain. Customize size, format, and theme to match your design needs. Accelerate design, prototyping, and content

NOVA-3951автор: NOVA-3951

Design Inspiration Server

Searches top design platforms like Dribbble and Behance to provide UI inspiration, color palettes, and layout patterns via the Serper API. It allows users to re

YonasValentinавтор: YonasValentin

PIX4Dmatic

Enables GUI automation for controlling PIX4Dmatic on Windows through MCP. Supports launching, focusing, capturing screenshots, sending hotkeys, clicking UI elem

jangjo123автор: jangjo123

Figma

Extract design specs and assets

Figmaавтор: Figma

mcp-dockmaster

An Open-Sourced UI to install and manage MCP servers for Windows, Linux and macOS.

автор: Community

ariekogan/ateam-mcp

Build, validate, and deploy multi-agent AI solutions on the ADAS platform. Design skills with tools, manage solution lifecycle, and connect from any AI environm

ariekoganавтор: ariekogan

thinkchainai/mcpbundles

MCP Bundles: Create custom bundles of tools and connect providers with OAuth or API keys. Use one MCP server across thousands of integrations, with programmatic

thinkchainaiавтор: thinkchainai

arikusi/nakkas

MCP server that turns AI into an SVG artist. One rendering engine with JSON config, AI controls all design parameters. CSS @keyframes + SMIL animations, 16+ ele

arikusiавтор: arikusi

Compare Openems with

Не уверен что выбрать?

Найди свой стек за 60 секунд

Автор?

Embed-бейдж для README

Похожее

Все в категории design