Math MCP Learning Server

Educational MCP server with 17 tools, persistent workspace, and cloud hosting. Built with FastMCP 2.0 and the official Model Context Protocol Python SDK.

Available on:

Official MCP Registry - io.github.clouatre-labs/math-mcp-learning-server
PyPI - math-mcp-learning-server

Requirements

Requires an MCP client:

Claude Desktop - Anthropic's desktop app
Claude Code - Command-line MCP client
Goose - Open-source AI agent framework
OpenCode - Open-source MCP client by SST
Kiro - AWS's AI assistant
Gemini CLI - Google's command-line tool
Any MCP-compatible client

Quick Start

Cloud (No Installation)

Connect your MCP client to the hosted server:

Claude Desktop (claude_desktop_config.json):

{
  "mcpServers": {
    "math-cloud": {
      "transport": "http",
      "url": "https://math-mcp.fastmcp.app/mcp"
    }
  }
}

Local Installation

Automatic with uvx (recommended):

{
  "mcpServers": {
    "math": {
      "command": "uvx",
      "args": ["math-mcp-learning-server"]
    }
  }
}

Manual installation:

# Basic installation
uv pip install math-mcp-learning-server

# With matrix operations support
uv pip install math-mcp-learning-server[scientific]

# With visualization support
uv pip install math-mcp-learning-server[plotting]

# All features
uv pip install math-mcp-learning-server[scientific,plotting]

Tools

Category	Tool	Description
Workspace	`save_calculation`	Save calculations to persistent storage
	`load_variable`	Retrieve previously saved calculations
Math	`calculate`	Safely evaluate mathematical expressions
	`statistics`	Statistical analysis (mean, median, mode, std_dev, variance)
	`compound_interest`	Calculate compound interest for investments
	`convert_units`	Convert between units (length, weight, temperature)
Matrix	`matrix_multiply`	Multiply two matrices
	`matrix_transpose`	Transpose a matrix
	`matrix_determinant`	Calculate matrix determinant
	`matrix_inverse`	Calculate matrix inverse
	`matrix_eigenvalues`	Calculate eigenvalues
Visualization	`plot_function`	Plot mathematical functions
	`create_histogram`	Create statistical histograms
	`plot_line_chart`	Create line charts
	`plot_scatter_chart`	Create scatter plots
	`plot_box_plot`	Create box plots
	`plot_financial_line`	Create financial line charts

Resources

math://workspace - Persistent calculation workspace summary
math://history - Chronological calculation history
math://functions - Available mathematical functions reference
math://constants/{constant} - Mathematical constants (pi, e, golden_ratio, etc.)
math://test - Server health check

Prompts

math_tutor - Structured tutoring prompts (configurable difficulty)
formula_explainer - Formula explanation with step-by-step breakdowns

See Usage Examples for detailed examples.

Development

# Clone and setup
git clone https://github.com/clouatre-labs/math-mcp-learning-server.git
cd math-mcp-learning-server
uv sync --extra dev --extra plotting

# Test server locally
uv run fastmcp dev src/math_mcp/server.py

Testing

# Run all tests
uv run pytest tests/ -v

# Run with coverage
uv run pytest tests/ --cov=src --cov-report=html --cov-report=term

# Run specific test category
uv run pytest tests/test_matrix_operations.py -v

Test Suite: 154 tests across 6 categories (Agent Card, HTTP Integration, Math, Matrix, Persistence, Visualization)

Code Quality

# Linting
uv run ruff check

# Formatting
uv run ruff format --check

# Security checks
uv run ruff check --select S

Security

The calculate tool uses restricted eval() with a whitelist of allowed characters and functions, restricted global scope (only math module and abs), and no access to dangerous built-ins or imports. All tool inputs are validated with Pydantic models. File operations are restricted to the designated workspace directory. Complete type hints and validation are enforced for all operations.

Math MCP Learning Server