By using this site, you agree to the Privacy Policy and Terms of Use.
Accept
Logic & LayersLogic & Layers
  • Tools
  • Earn with AI
  • Productivity
  • Automation
  • Guides
Logic & LayersLogic & Layers
  • Privacy Policy
  • About
Search
  • Tools
  • Earn with AI
  • Productivity
  • Automation
  • Guides
  • About
  • Contact
  • Blog
  • Privacy Policy
  • Complaint
  • Advertise
© 2026 Logic and Layers. Ruby Design Company. All Rights Reserved.
Claude Code terminal showing KiCad MCP server generating PCB layout
Tools

Claude Code for KiCad: Automate PCB Design with AI Scripts

Editorial Team
Last updated: August 22, 2026 7:19 pm
Editorial Team
Share
Claude Code with KiCad MCP automating PCB design

You’ve spent hours placing components in KiCad, routing traces by hand, and praying the DRC passes. What if you could describe a board once and generate ten variants before lunch? That’s not a future promise. It works right now.

Contents
The Problem: KiCad’s Steep Learning CurveThe Solution: Claude Code + MCP = Programmatic PCB DesignWhat is Claude Code (vs Claude Desktop)What is MCP and KiCAD-MCP-ServerSetup: Get Running in 15 MinutesPrerequisitesInstall KiCAD-MCP-ServerConfigure Claude Code MCPExample 1: Generate a Basic MCU BoardExample 2: Parameterized Board VariantsExample 3: Automated ERC/DRC Validation in CI/CDPro Tips: Going FurtherParts Registry IntegrationKonnect: The Rust Native FutureCustom Footprint GenerationTakeaway

Most people use Claude by chatting with it. You describe a circuit, it spits out KiCad Python code, you copy-paste, hope it runs. That’s fine for one-offs. But it doesn’t scale. The real power sits in Claude Code — Anthropic’s CLI agent — paired with an MCP server that turns KiCad into a programmable toolkit.

KiCad’s new Python API (since 9.0) is powerful but poorly documented. The SWIG bindings are fragile. Learning pcbnew from scratch takes weeks. The KiCAD-MCP-Server bridges all of it. You write a prompt. Claude Code calls the MCP tools. KiCad executes. You get a board file, not a chat transcript.

The Problem: KiCad’s Steep Learning Curve

KiCad is free, open source, and capable. It’s also hostile to newcomers. The UI demands you learn footprint libraries, netlist workflows, zone filling, design rule configuration — before you route a single trace. The Python API lets you automate everything. But the documentation reads like a reference manual written by engineers for engineers. No tutorials. No examples. Just class hierarchies.

You can script KiCad. Almost nobody does because the barrier is too high.

I know this because I tried. Spent three weekends fighting pcbnew’s Python bindings. The SWIG layer crashes on valid calls. The type hints are wrong. The examples in the repo target KiCad 7.0 and don’t work on 9.0. By Sunday night I had a script that placed one resistor. That’s not automation. That’s a party trick.

The real problem isn’t KiCad. It’s the gap between what the API can do and what the docs show you how to do.

The Solution: Claude Code + MCP = Programmatic PCB Design

What is Claude Code (vs Claude Desktop)

Claude Desktop is a chat app. You type, it replies. Claude Code is a coding agent that lives in your terminal. It reads files, runs commands, writes code, and — critically — uses MCP servers to extend its reach. You don’t paste code from a chat window. You tell it what you want and it modifies your project directly.

What is MCP and KiCAD-MCP-Server

MCP (Model Context Protocol) is an open standard from Anthropic. It defines how AI tools talk to external systems. The KiCAD-MCP-Server (maintained by mixelpixx, 1.9k stars on GitHub) exposes KiCad’s Python API as MCP tools. Claude Code calls these tools. KiCad does the work.

The server handles the messy SWIG bindings. You get clean operations like create_board, add_component, route_trace, run_drc. No boilerplate.

Setup: Get Running in 15 Minutes

Prerequisites

  • KiCad 9.0 or higher (the Python API changed significantly)
  • Python 3.10+
  • uv package manager (pip install uv)
  • Claude Code installed and authenticated
  • Git (for version control)

Install KiCAD-MCP-Server

git clone https://github.com/mixelpixx/KiCAD-MCP-Server
cd KiCAD-MCP-Server
uv sync

The server includes a TypeScript MCP layer and a Python backend that talks to KiCad. It’s MIT licensed and actively maintained.

Configure Claude Code MCP

Create or edit ~/.claude/mcp_servers.json:

{
  "mcpServers": {
    "kicad": {
      "command": "uv",
      "args": ["run", "server.py"],
      "cwd": "/path/to/KiCAD-MCP-Server"
    }
  }
}

Restart Claude Code. Run /mcp in the CLI to verify the server connects. You should see KiCad tools listed.

Example 1: Generate a Basic MCU Board

Open a terminal in your project folder. Start Claude Code:

claude

Now prompt it:

Create a KiCad project for a basic STM32F401 breakout board. Two layers. Include USB-C power, 8MHz crystal, SWD header, and 0.1″ headers for all GPIOs. Save as stm32_breakout.kicad_pro.

Claude Code will:
1. Create the project structure
2. Generate the schematic with proper symbols
3. Assign footprints (USB-C, QFP48, 0603 passives, pin headers)
4. Create the PCB outline
5. Place components logically
6. Route power and critical signals
7. Fill ground zones
8. Run DRC

You get a complete, manufacturable board. Not a suggestion. A .kicad_pcb file you can send to JLCPCB.

The first time I ran this, I expected a mess. I got a clean two-layer board with proper decoupling caps on every power pin, a solid ground pour, and zero DRC violations. The SWD header was oriented correctly for a J-Link. The USB-C footprint had the CC resistors populated. This isn’t toy output.

Example 2: Parameterized Board Variants

Here’s where it gets interesting. You don’t want one board. You want a family.

Create a parameterized script that generates STM32 breakout variants: STM32F401, STM32F411, STM32G431. Each variant should adjust the footprint, pin mapping, and crystal frequency automatically. Output separate KiCad projects in variants/.

Claude Code writes a Python script using the MCP tools. Run it once. Get three production-ready projects. Change the parameter list. Get ten. This is impossible with the chat interface.

Example 3: Automated ERC/DRC Validation in CI/CD

KiCad 9.0 includes kicad-cli for headless operation. The MCP server wraps it.

Add a GitHub Actions workflow that runs ERC and DRC on every push. Fail the build if violations exist. Use the MCP server to load the board and execute validation.

The resulting workflow:

name: PCB Validation
on: [push, pull_request]
jobs:
  drc:
    runs-on: ubuntu-latest
    container: kicad/kicad:9.0
    steps:
      - uses: actions/checkout@v4
      - name: Run ERC/DRC
        run: |
          kicad-cli sch erc *.kicad_sch --exit-code-violations
          kicad-cli pcb drc *.kicad_pcb --exit-code-violations

Your PCB design gets the same rigor as your code.

Pro Tips: Going Further

Parts Registry Integration

The Seeed Studio fork of KiCAD-MCP-Server includes PartReel — 21,000+ verified footprints and symbols. No more hunting SnapEDA or drawing your own.

Search PartReel for “USB-C 16pin” and add the recommended footprint to my board.

Konnect: The Rust Native Future

The same team is building Konnect — a native KiCad 10 plugin written in Rust. 171 tools. No Python SWIG layer. Bundled Claude skills. Design-review audits. Manufacturing pipeline. It’s AGPL-3.0 (free for individuals, commercial licenses for businesses). The Python server stays MIT and maintained. Konnect is where new development happens.

Custom Footprint Generation

Need a weird footprint? Datasheet in hand?

Generate a KiCad footprint for the RP2040 QFN-56 package from the official datasheet dimensions. Include courtyard, silkscreen, and 3D model placeholder.

Claude Code reads the PDF, extracts dimensions, writes the .kicad_mod file. Done.

Takeaway

Stop chatting with AI about hardware. Start scripting it. Clone KiCAD-MCP-Server. Configure Claude Code. Describe your next board as a parameterized template, not a one-off conversation. You’ll never place a footprint by hand again.

The tools are free. The learning curve just collapsed. Your next PCB is waiting.

You Might Also Like

AI Agent Security: Complete Guide to Securing Your AI Tools in the Cloud (2026)
AI layoffs are real — but so is the hype: The automation reality check beginners need
Local LLM Hardware Guide 2026: DDR5 Prices Up 500% – What to Buy Instead
Best AI website builders in 2026 (Free tools that actually work)
Open-Weight AI Models: GLM 5.3 and What It Means for You
TAGGED:AI CodingClaude Codehardware designKiCadPCB automation
Share
Previous Article PCB design with Claude AI using KiCad-MCP PCB Design with Claude: AI-Assisted Hardware Design for Beginners
Next Article Diagram showing adversarial perturbation in AI models - visual representation of security vulnerabilities in artificial intelligence systems AI Agent Security: Complete Guide to Securing Your AI Tools in the Cloud (2026)
Leave a Comment

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

banner banner
Create an Amazing Newspaper
Discover thousands of options, easy to customize layouts, one-click to import demo and much more.
Learn More

Latest News

The Most Customizable LLM Chat App Is Free and Open Source (Setup Guide)
Tools
Hooded hacker figure with the OpenAI logo as a face, surrounded by panicked emoji faces on a blue and orange background
OpenAI’s Hugging Face Hack: What It Means for Your AI Safety
Tools
Smartphone displaying the Claude app logo with the orange Anthropic starburst icon on a black background
Claude Code Session Messaging + Auto Mode: Complete Beginner Guide
Productivity
How Headway Built Custom AI Tool with Claude Code SDK
Guides

Recent Posts

  • The Most Customizable LLM Chat App Is Free and Open Source (Setup Guide)
  • OpenAI’s Hugging Face Hack: What It Means for Your AI Safety
  • Claude Code Session Messaging + Auto Mode: Complete Beginner Guide
  • How Headway Built Custom AI Tool with Claude Code SDK
  • Gemini Chrome Select from Screen: Beginner Guide

Recent Comments

  1. I tested 6 AI task managers for 30 days (Only 3 are worth it) on Best AI time blocking apps in 2026 (I tested 5 that survive when your schedule falls apart)
  2. Gemini CLI: How to Start Coding with AI for Free on How to use Google Gemini 3.5 Flash Search: A complete beginner guide
  3. GitHub Copilot's New Pricing: 10x More Expensive? | Logic & Layers on Cancel ChatGPT, Perplexity & Gemini — use Claude instead
  4. Google Gemini Spark Review: Is It Worth Using? | Logic & Layers on Gemini in Android Auto: Complete beginner’s guide (2026)
  5. Google Gemini Spark Review: Is It Worth Using? | Logic & Layers on Cancel ChatGPT, Perplexity & Gemini — use Claude instead

You Might also Like

NVIDIA PersonaPlex running as a local AI speech model on a laptop screen
Tools

NVIDIA PersonaPlex: how to run this free AI speech model locally

Editorial Team
Editorial Team
9 Min Read
Visual workflow builder connecting trigger events to automated action sequences
Tools

How to build your first AI automation workflow in 20 minutes

Zero
Zero
16 Min Read
Automation

Build an AI Code Review Bot in 30 Minutes with Vercel Eve

Editorial Team
Editorial Team
9 Min Read
//

We influence 20 million users and is the number one business and technology news network on the planet

Quick Link

  • PRIVACY NOTICE
  • YOUR PRIVACY RIGHTS
  • INTEREST-BASE ADSNew
  • TERMS OF USE
  • OUR SITE MAP

Support

  • ADVERTISE
  • ONLINE BESTHot
  • CUSTOMER
  • SERVICES
  • SUBSCRIBE

Categories

  • Tools
© 2026 Logic and Layers. All Rights Reserved.