Cruncher Mcp

@islobodan

About Cruncher Mcp

Cruncher MCP: The Scientific Calculator MCP Server

#calculator-math-scientific-statistics-unit-conversion-model-context-protocol-model-context-protocol-server-zero-dependencies

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "cruncher": {
      "command": "npx",
      "args": [
        "-y",
        "@slbdn/cruncher-mcp"
      ],
      "env": {
        "CRUNCHER_TOOL_SET": "standard",
        "CRUNCHER_TIMEOUT": "5000"
      }
    }
  }
}

Tools

Adds two numbers.

Subtracts two numbers.

Multiplies two numbers.

Divides two numbers. Errors on zero divisor.

Raises a to the power of b.

Square root. Errors on negative input.

Sine. Angle in degrees by default, or radians with unit: "radians".

Cosine. Angle in degrees by default, or radians with unit: "radians".

Tangent. Angle in degrees by default, or radians with unit: "radians".

Arcsine. Result in degrees by default, or radians with unit param.

Arccosine. Result in degrees by default, or radians with unit param.

Arctangent. Result in degrees by default, or radians with unit param.

Base-10 logarithm. Errors on non-positive input.

Natural logarithm (ln). Errors on non-positive input.

Absolute value.

Returns a mathematical, physical, or chemical constant. See enum values.

Sum of numbers.

Average (mean) of numbers.

Median of numbers.

Minimum of numbers.

Maximum of numbers.

Count elements.

Range (max - min) of numbers.

Sample variance (n-1). Set population: true for population variance (n).

Sample standard deviation (n-1). Set population: true for population (n).

What is X% of Y? e.g., 15% of 200 = 30

Percentage change from A to B. e.g., 50→80 = +60%

X is Y% of what? e.g., 30 is 15% of 200

Remainder of a / b. Errors on zero divisor.

Factorial of non-negative integer (n!). n > 170 overflows.

Evaluate a mathematical expression. PRIMARY tool for ALL math: +, -, *, /, %, ^. Functions: sqrt, sin, cos, tan, asin, acos, atan, log10, ln, log(x,base), abs, round, floor, ceil, min, max. Constants: pi, e, tau, phi, sqrt2, euler_mascheroni, c, g, G, h, k, R, NA, e_charge, m_e, m_p. Parentheses and scientific notation (1e6) supported. Use explicit operators: 2 * pi, not 2pi.

Convert between common units. Categories: length, weight, temperature, area, volume, time, speed, digital_storage.

Set global trig angle mode. Individual calls with unit param override this.

Get current trig angle mode.

Overview

What is Cruncher Mcp?

Cruncher Mcp is a powerful scientific calculator built as a Model Context Protocol (MCP) server that gives AI assistants the ability to perform complex mathematical calculations, run statistical analyses, manage memory, and access scientific constants. It is designed with zero external dependencies, relying solely on native Node.js capabilities, and works with any MCP-compatible client such as Claude Desktop, Cursor IDE, OpenWebUI, and others.

How to use Cruncher Mcp?

Install and run the server via npx @slbdn/cruncher-mcp, then configure your MCP client (e.g., Claude Desktop) to connect to it by adding a new “cruncher” server with the command npx and arguments -y @slbdn/cruncher-mcp. Alternatively, download cruncher.js from GitHub and configure the client with a node command pointing to the file. The tool tier can be set via the environment variable CRUNCHER_TOOL_SET (minimal, standard, or full), and the calculation timeout can be adjusted with CRUNCHER_TIMEOUT (milliseconds, default 3000).

Key features of Cruncher Mcp

Zero‑dependency scientific calculator for AI assistants
Strict input validation with custom schema validator
Infinite loop protection using Node.js worker_threads (3‑second timeout)
Accurate decimal math using integer‑scaling logic
Three configurable tool tiers (minimal, standard, full)
80+ unit conversions across 8 categories
Memory functions: M+, M‑, MR, MC (full tier only)
Scientific constants: Math, Physics, and Chemistry values

Use cases of Cruncher Mcp

Let your AI assistant evaluate complex mathematical expressions during a conversation
Compute statistical measures (average, median, variance, standard deviation) for data sets
Convert between units of length, weight, temperature, volume, and more
Perform percentage calculations, angle trigonometry, and number‑theoretic operations

FAQ from Cruncher Mcp

What is the Model Context Protocol (MCP)?

MCP is an open standard that allows AI applications to securely connect with external data sources and tools—a universal API for AI. Cruncher implements this protocol so any MCP‑capable AI can use it without custom integration.

Does Cruncher Mcp require any external libraries?

No. Cruncher is built with zero dependencies, relying purely on Node.js standard libraries. It uses new Function() with a strict regex whitelist for expression evaluation, not eval() or libraries like mathjs.

What runtime and environment variables are needed?

Node.js version 18.0.0 or newer is required. The server supports two environment variables: CRUNCHER_TOOL_SET (values: minimal, standard, full) and CRUNCHER_TIMEOUT (timeout in milliseconds, default 3000).

How does Cruncher Mcp protect against infinite loops or heavy calculations?

It spawns calculations in a Node.js worker_thread and terminates the worker if it exceeds a configurable timeout (default 3 seconds), keeping the main server responsive.

What transport and authentication does Cruncher Mcp use?

Cruncher communicates over stdio using JSON‑RPC messages (initialize, tools/list, tools/call). The README does not mention any authentication mechanism or network transport.

Frequently asked questions

What is the Model Context Protocol (MCP)?

Does Cruncher Mcp require any external libraries?

No. Cruncher is built with zero dependencies, relying purely on Node.js standard libraries. It uses `new Function()` with a strict regex whitelist for expression evaluation, not `eval()` or libraries like `mathjs`.

What runtime and environment variables are needed?

Node.js version 18.0.0 or newer is required. The server supports two environment variables: `CRUNCHER_TOOL_SET` (values: minimal, standard, full) and `CRUNCHER_TIMEOUT` (timeout in milliseconds, default 3000).

How does Cruncher Mcp protect against infinite loops or heavy calculations?

It spawns calculations in a Node.js `worker_thread` and terminates the worker if it exceeds a configurable timeout (default 3 seconds), keeping the main server responsive.

What transport and authentication does Cruncher Mcp use?

Cruncher communicates over `stdio` using JSON‑RPC messages (initialize, tools/list, tools/call). The README does not mention any authentication mechanism or network transport.