brainycalc

The trusted computational backend for AI agents.

BrainyCalc is a Model Context Protocol (MCP) math server that gives LLMs and AI agents accurate, independently verified, and reproducible mathematical results — with full auditability.

Instead of letting an LLM estimate mathematical or scientific answers, the AI sends computations to BrainyCalc, which handles the heavy lifting with precision and verification.

| Metric | Value | |---|---| | Reproducibility | 100% — same input, same output | | Error bound | ±10⁻¹⁵ per computation | | Verification | 6-step check on every result | | Audit trail | Full cryptographic proof |

Quick Start

Add the following to your MCP configuration file:

{
  "mcpServers": {
    "brainycalc": {
      "command": "npx",
      "args": ["-y", "brainycalc"]
    }
  }
}

Available Tools

BrainyCalc exposes 100+ tools across four categories.

Mathematics (math_*)

| Tool | Description | |---|---| | math_evaluate | Safely evaluate an arithmetic expression string | | math_percentage | Percentage of, percentage change, percentage difference | | math_statistics | Mean, median, mode, std dev, variance for a dataset | | math_quadratic | Solve ax² + bx + c = 0 (real and complex roots) | | math_gcd_lcm | Greatest Common Divisor and Least Common Multiple | | math_prime_factors | Prime factorization and primality check | | math_power_root | Powers and nth roots | | math_permutations_combinations | Factorial, nPr (permutations), nCr (combinations) | | math_simplify_fraction | Reduce fraction to lowest terms, convert to mixed number | | math_geometry | Area, perimeter, volume for circle, rectangle, triangle, sphere, cylinder | | math_trigonometry | sin, cos, tan, asin, acos, atan (degrees or radians) | | math_logarithm | Logarithm with any base |

Unit Conversion (convert_*)

80+ converters spanning every major physical quantity:

• Everyday: length, mass, volume, temperature, speed, area, time
• Electrical: voltage, current, resistance, capacitance, inductance, charge, conductance, resistivity, field strength
• Mechanics: force, pressure, energy, power, torque, acceleration, momentum, density, flow rate
• Thermodynamics: heat flux, thermal conductivity, specific heat capacity, thermal expansion
• Optics & Radiation: illumination, luminance, luminous intensity, wavelength ↔ frequency, radiation dose, radiation exposure
• Data: storage (bits to petabytes), transfer rates, image resolution
• Chemistry: molar/solution concentration, viscosity (dynamic & kinematic)
• Engineering: angular velocity, angular acceleration, moment of inertia, surface tension
• Specialist: typography (pt/px/em), number base (binary, hex, octal), SI prefix scaling

Fitness & Health (fitness_*)

| Tool | Description | |---|---| | fitness_bmi | Body Mass Index with WHO category | | fitness_bmr | Basal Metabolic Rate (Mifflin-St Jeor or Harris-Benedict) | | fitness_tdee | Total Daily Energy Expenditure from BMR + activity level | | fitness_body_fat | US Navy body fat percentage formula | | fitness_ideal_weight | Ideal body weight (multiple formulas) | | fitness_macros | Macronutrient targets from calorie goal | | fitness_target_heart_rate | Target heart rate zones by age and intensity | | fitness_water_intake | Daily water intake recommendation |

Date & Time (datetime_* / calc_*)

| Tool | Description | |---|---| | datetime_difference / calc_date_diff | Days, hours, minutes between two dates | | datetime_add_subtract / calc_time_add | Add or subtract time from a date | | datetime_calculate_age / calc_age | Precise age from date of birth | | datetime_day_of_week / calc_day_of_week | Day name for any date | | datetime_work_days | Working days between two dates | | datetime_unix_timestamp | Convert between ISO dates and Unix timestamps | | calc_hours_between | Hours and minutes between two times | | calc_time_duration | Duration formatting | | calc_world_time / calc_local_time | Time zone conversions | | calc_day_counter | Count days to/from a target date | | calc_due_date | Estimated due date calculator | | calc_pregnancy | Pregnancy week and trimester from LMP | | calc_ovulation | Ovulation and fertility window | | calc_period | Menstrual cycle tracking |

Why BrainyCalc?

LLMs are powerful reasoners but unreliable calculators. Floating-point arithmetic, unit conversion chains, and multi-step statistical computations are common sources of silent errors in AI-generated answers. BrainyCalc offloads these to a deterministic engine that:

• Never estimates — uses reliable numerical or symbolic methods
• Verifies independently — a second pass catches instabilities before results reach your agent
• Is fully reproducible — same input always produces same output; non-determinism is eliminated
• Reports confidence — every result includes a confidence score and error bound
• Records provenance — engine version, method, and parameters are captured with each result
• Maintains an audit trail — fingerprinted results can be inspected or reproduced at any time

This makes BrainyCalc suitable for use cases where accuracy is non-negotiable: healthcare calculations, financial modelling, engineering, scientific research, legal and compliance workflows.

Integration

BrainyCalc speaks standard MCP over JSON-RPC stdio, so it works with any MCP-compatible platform:

• Claude Desktop (Anthropic)
• LangChain agents
• Custom SDK implementations
• Self-hosted or cluster deployments (locally runnable for lower latency)

Tool schemas are automatically exposed via the MCP tools/list endpoint — no manual wiring required.

Requirements

• Node.js ≥ 18
• npm ≥ 7

License

MIT

Links

• Homepage: brainycalc.com
• How it works: brainycalc.com/#how-it-works
• Use cases: brainycalc.com/#usecases