mcpField
N
ngspice-mcp

ngspice-mcp

Local
github.com/gtnobleScience & Education19Updated 80w ago
Repository

About ngspice-mcp

ngspice-mcp exposes ngspice circuit simulation through an MCP server for AI language models and other MCP clients. It supports loading netlists, running simulations, and retrieving plotted vector data.

Key features

  1. 01Load circuits and netlists
  2. 02Run ngspice simulations
  3. 03List plots and vectors
  4. 04Retrieve vector data in multiple formats
  5. 05Read stdout and stderr resources

Use cases

  1. 01Load a circuit netlist for simulation
  2. 02Run DC, AC, or transient analyses
  3. 03Retrieve selected simulation vectors and intervals

MCP support

Provides MCP tools for circuit loading, simulation, and data retrieval, along with stdout, stderr, and usage resources.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

No install method listed yet

We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.

Repository

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

ngspice-mcp

A Model Context Protocol (MCP) server that provides access to ngspice circuit simulation functionality through a standardized protocol interface. This server enables AI language models to interact with ngspice in a controlled and structured way.

Features

  • Integration with ngspice's shared library interface
  • Standardized MCP protocol implementation
  • Synchronous operation with structured data access
  • Comprehensive output capture and redirection
  • Vector data handling with multiple representation formats
  • Support for both server and library usage modes
  • Built-in test suite

Requirements

  • D compiler (DMD/LDC)
  • ngspice shared library
  • MCP server library (d-mcp-server)
  • D build system (dub)

Installation

  1. Ensure system requirements are installed
  2. Clone the repository
  3. Build using dub:
dub build --config=server

Usage

The server can be started with:

./ngspice-mcp [options]
Command Line Options
  • --working-dir, -d: Set the working directory for circuit files (default: current directory)
Available Tools
Circuit Loading
  • loadCircuit: Load circuit netlists directly
  • loadNetlistFromFile: Load netlists from files

Example netlist:

RC Circuit
R1 in out 1k
C1 out 0 1u
.end
Simulation
  • runSimulation: Execute simulation commands

Common commands:

  • op: DC operating point
  • dc source start stop step
  • ac dec points fstart fend
  • tran step tstop
Data Access
  • getPlotNames: List available simulation plots
  • getVectorNames: List vectors in a specific plot
  • getVectorData: Retrieve vector data with options for:
    • Magnitude-phase representation
    • Rectangular (real-imaginary) representation
    • Both representations
    • Optional interval selection
Resources
  • stdout://: Standard output stream from ngspice
  • stderr://: Error output stream from ngspice
  • usage://: Comprehensive usage guide

Development

Project Structure
ngspice-mcp/
├── source/
│   ├── app.d           # Main application
│   ├── bindings/       # ngspice C API bindings
│   ├── database/       # SQLite database handling
│   ├── parser/         # Netlist parsing
│   └── server/         # MCP server implementation
├── resources/          # Resource files
└── bin/               # Build outputs
Build Configurations
  1. Server Mode
dub build --config=server
  1. Library Mode
dub build --config=library
  1. Unit Tests
dub test --config=unittest
Testing

Run the test suite:

dub test

Tests cover:

  • ngspice bindings
  • Server functionality
  • Tool validation
  • Resource handling
  • Error cases

Technical Details

Architecture

The server implements several key components:

  1. NgspiceServer: Core server implementation

    • Tool registration and handling
    • Resource management
    • Output stream capture
    • Vector data processing

  2. Output System

    • Stdout/stderr capture
    • Stream buffering
    • Resource notification

  3. Vector Processing

    • Complex number handling
    • Scientific notation formatting
    • Interval selection
    • Multiple representation formats
Database System
  • Model parameter storage
  • Efficient indexing and querying
  • Transaction support
  • Concurrent access handling

License

MIT License - see dub.json for details.

Copyright © 2024, Garret Noble

Contributing

Contributions are welcome! Please ensure:

  1. Code follows project style and standards
  2. Tests are included for new functionality
  3. Documentation is updated as needed
  4. Commit messages are clear and descriptive

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.