KiCAD

This KiCAD MCP Server provides a bridge between AI assistants and KiCAD, enabling natural language control of PCB design tasks via the MCP protocol. The server exposes a catalog of 64 tools with JSON Schema validation, organized by a router pattern to keep AI context lean while still giving access to all capabilities. It can interact with KiCAD projects in real time, including creating projects, placing components, wiring nets, and loading dynamic symbols from KiCAD libraries. The JLCPCB integration adds parts search and pricing, helping select components directly within a workflow. To use, start the MCP server and connect your AI agent to the MCP endpoint; then request actions like creating schematics, placing components, routing wires, or exporting Gerber/NPC files. The router helps Claude or similar agents discover the right tool without loading every tool schema into memory, while still allowing explicit tool selection when needed.

Typical workflows include: (1) loading dynamic symbols from KiCAD symbol libraries, (2) placing and wiring components with automatic pin rotation and routing strategies, (3) querying real-time project state, and (4) selecting parts from JLCPCB catalogs with pricing and stock checks. You can also enable IPC (experimental) for live UI synchronization with KiCAD, so changes appear immediately in the KiCAD UI during a session.

Prerequisites:

• A supported environment (Linux, macOS, or Windows) with Git and Node.js installed if using the Node/npx approach.
• Basic KiCAD installation for local project creation and manipulation.
• Optional: IPC API enabled in KiCAD for real-time UI synchronization.

Install steps (npx-based approach):

1. Ensure Node.js and npm are installed:
   • macOS/Linux: curl -fsSL https://deb.nodesource.com/setup_current.x | sudo -E bash - && sudo apt-get install -y nodejs
   • Windows: install Node.js from https://nodejs.org/
2. Run the MCP server via npx (no global install required):
   • npx -y mixelpixx-kicad-mcp-server
3. If you prefer a persistent install, you can install the package locally and run it from a script:
   • npm install --save-dev mixelpixx-kicad-mcp-server
   • node node_modules/mixelpixx-kicad-mcp-server/server.js
4. (Optional) Docker deployment:
   • docker run -i mixelpixx/kicad-mcp-server
5. Verify connectivity by connecting your MCP client to the server endpoint and listing available tools or category tools.

Notes:

• Ensure KiCAD is installed and accessible for project creation and manipulation.
• If you plan to use JLCPCB features, ensure any required API keys or credentials are supplied as environment variables per your deployment method.

Tips and common issues:

• If the MCP client reports missing tools, rely on the router by invoking list_tool_categories or get_category_tools before execute_tool to reduce context load.
• When enabling IPC (experimental), ensure KiCAD is configured to allow IPC API Server access and that the firewall permits the IPC port.
• For dynamic symbol loading, make sure KiCAD symbol libraries are available on disk; the server can load from standard KiCad library formats (.kicad_sym).
• Check logs for tool execution errors and version compatibility with KiCAD project files (.kicad_pcb, .kicad_sch).
• If you encounter authentication or pricing errors with JLCPCB integration, verify local library search paths and ensure API endpoints are reachable.
• Environment variables you may consider adding: MCP_KICAD_IPC_PORT, MCP_KICAD_LIB_PATH, JLCPCB_API_KEY, KICAD_PROJECT_ROOT.