Run ngspice circuit simulations in batch mode and retrieve parsed measurement results as JSON.
spice_run — Simulate a .cir/.sp netlist with ngspice -b, returning parsed .meas values (e.g. {name, value}) and a short log tail; no plots.
netlist (required): path to the netlist file.
cwd (optional): working directory.
timeout_ms (optional): max runtime, default 60000 ms.
spice_toolchain_info — Probe the engine: returns the ngspice version and the active container/host runtime (e.g. rootless Podman image).
Intended for AI coding agents and IDEs (Cursor, Windsurf, Copilot/Codex, Claude Code, etc.) via MCP stdio, so agents can iterate on circuit designs without scrolling raw simulator logs.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-spice Simulate rc_lowpass.cir and return the .meas results as JSON"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Run ngspice circuit netlists and retrieve measurement results through an MCP server.
Who it is for: Hardware engineers using an MCP-capable client or coding agent.
First task: Configure the server in your MCP client, then call spice_toolchain_info before running a design.
What to expect: Simulator availability, then parsed circuit measurements and execution status.
Current scope: Published MCP server. The npx command starts a stdio server that waits for a client; EDA execution also needs its documented host/container tools.
Start here: Runtime requirements and configuration.
Related projects: eda-docker-images, hw-agent-scaffold.
Model Context Protocol (MCP) server for ngspice batch circuit simulation.
mcp-spice lets AI coding agents and IDEs (Cursor, Windsurf, GitHub Copilot / OpenAI Codex, Claude Code, Google Antigravity, OpenCode, Cline) run SPICE netlists with ngspice in batch mode and get .meas results back as JSON instead of plot windows or thousand-line logs.
One-step with the rest of the family:
npx -y @zesun33/create-hw-agent my-asicOr run this published server alone:
npx -y @zesun33/mcp-spiceRelated MCP server: ltagent-mcp
Tool | Parameters | Engine | Description |
|
|
| Batch simulate a |
| none | Probe | ngspice version and active runtime/image. |
// Tool Call: spice_run {"netlist": "rc_lowpass.cir"}
{
"success": true,
"measurements": [{ "name": "vmid", "value": 0.63 }]
}mcp-spice runs inside the zesun33/spice rootless Podman image.
Public install (recommended — anyone can pull):
podman pull ghcr.io/zesun33/spice:latest
export MCP_SPICE_IMAGE=ghcr.io/zesun33/spicePublic GHCR (ghcr.io/zesun33/spice) is the default. Local builds still work as localhost/zesun33/spice via MCP_SPICE_IMAGE. To force host binaries: export MCP_SPICE_RUNTIME=host.
Environment | Setup Location |
AI IDEs |
|
CLI Agents | Claude Code / OpenCode MCP config |
Desktop |
|
{
"mcpServers": {
"spice": {
"command": "npx",
"args": ["-y", "@zesun33/mcp-spice"],
"env": { "MCP_SPICE_IMAGE": "ghcr.io/zesun33/spice" }
}
}
}Until the package is on npm, point command at node and args at this repo's dist/index.js.
./scripts/verify.sh # full, needs podman + spice image
./scripts/verify.sh --quick # CISee RELEASING.md for GitHub Actions dry runs and trusted publishing.
Runs a SPICE netlist in ngspice batch mode (-b) and returns parsed .meas results plus a short log tail. Does not plot. Use spice_toolchain_info to probe the engine.
Inputs: netlist (required), cwd, timeout_ms.
Returns active container/host runtime and the ngspice version string.
Inputs: no named parameters.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 2 tools
spice_run executes a simulation and returns results, while spice_toolchain_info reports environment/version details. Their purposes are completely distinct with no functional overlap.
Both tools share the spice_ prefix and snake_case style, but spice_run uses a verb while spice_toolchain_info is more noun-like. This is a minor inconsistency rather than a chaotic naming scheme.
With only two tools, the set is slightly below the typical 3-15 range, but it is still reasonable for a narrowly focused SPICE execution server. The small count matches the minimal scope.
The server covers the core run-and-inspect workflow with parsed .meas results and log tail access. It lacks raw output or full-log retrieval, but these are not obvious dead ends for the intended simple simulation use case.
Documentation for the Spektralwerk spectrometer SCPI API as a streamable HTTP MCP Server
Design, route, check, review and export manufacturable PCBs from plain language; KiCad import.
A free quantum job service: OpenQASM 3 run, checked and signed, with no account or key.
Remote MCP: JSON repair, JSON Schema validation, JSON to TypeScript, CSV to JSON and JSON to CSV.
