spicebridge

SPICEBridge is an MCP server that provides AI-assisted circuit design and SPICE simulation via ngspice. You install the Python package spicebridge and run the server through the spicebridge command. The server exposes a suite of tools for creating, configuring, simulating, measuring, and exporting SPICE circuits, including templates, netlist generation, circuit composition, Monte Carlo analyses, and model generation. In practice, you describe a circuit in plain language, let the server generate a netlist, run simulations (AC, transient, DC operating point), perform measurements (bandwidth, gain, DC levels, transient metrics, power), and verify results against target specs. You can also export KiCad schematics, render interactive web viewers, and use a cloud deployment wizard to expose the service publicly via a named tunnel or cloud-based endpoint.

To use it, run the server in your environment and interact with its tools through the API exposed by the spicebridge command. The core workflow spans creating or loading templates, configuring component values, running simulations, extracting results, and evaluating them against design criteria. Tools are organized into categories such as Create & Configure, Simulate, Measure, Evaluate & Export, Composition & Ports, and Advanced Analysis, each exposing commands like load_template, run_ac_analysis, measure_bandwidth, draw_schematic, set_ports, and run_monte_carlo. You can also generate SPICE models from datasheets or export KiCad schematics for downstream PCB design.

If you plan to expose the server over the web, use the Cloud setup wizard (spicebridge setup-cloud) to provision a tunnel and domain, and then share the connection information with your AI workflows.

Prerequisites:

• Python 3.10 or later
• ngspice installed and available on your system PATH
• Internet access to install Python packages

1. Create and activate a Python virtual environment (recommended): python -m venv venv source venv/bin/activate # Linux/macOS venv\Scripts\activate # Windows

2. Install SPICEBridge from PyPI: pip install spicebridge

3. Verify ngspice is available in PATH: ngspice -v

4. Run the SPICEBridge server locally: spicebridge

5. Optional: Run with cloud deployment if you plan to expose it publicly: spicebridge setup-cloud

6. (Optional) Install development dependencies and run tests: git clone https://github.com/clanker-lover/spicebridge.git cd spicebridge pip install -e ".[dev]" pytest

Notes and tips:

• NGspice is required for SPICE simulations; ensure it is installed and reachable from your shell before starting spicebridge.
• The local server defaults to a minimal configuration; modify mcp.json in your project to wire spicebridge into your MCP workflow (e.g., include in a Claude-compatible config).
• For distributed usage, the Cloud setup wizard handles automatic cloudflared installation, authentication, and tunnel management. Named tunnels provide persistent domains; quick tunnels are temporary.
• When using the web-based viewer or KiCad export features, ensure your environment has graphical/tooling support or appropriate browser access for the viewer and download features.
• If you encounter port or DNS issues during cloud setup, review your network firewall, ensure the chosen port is open, and confirm domain DNS propagation if using a custom domain.
• You can use the various tools by name (e.g., load_template, run_ac_analysis, measure_bandwidth, draw_schematic, export_kicad) to compose complex design workflows within your MCP integrations.