mcp-sympy

MCP server that exposes SymPy's symbolic mathematics functionality

Install

pip install mcp-sympy

MCP Server

mcp-name: io.github.daedalus/mcp-sympy

Usage

As MCP Server

Configure in your MCP client:

{
  "mcpServers": {
    "mcp-sympy": {
      "command": "mcp-sympy",
      "env": {}
    }
  }
}

Python API

from mcp_sympy import tools

# Simplify expressions
result = tools.sympy_simplify("sin(x)**2 + cos(x)**2")
print(result)  # "1"

# Solve equations
result = tools.sympy_solve("x**2 - 4", "x")
print(result)  # "[2, -2]"

# Differentiate
result = tools.sympy_diff("x**2", "x")
print(result)  # "2*x"

# Integrate
result = tools.sympy_integrate("x**2", "x")
print(result)  # "x**3/3"

# Create matrices
result = tools.sympy_matrix("1,2; 3,4")
print(result)  # "Matrix([[1, 2], [3, 4]])"

# Number theory
result = tools.sympy_isprime("7")
print(result)  # "True"

Available Tools

The MCP server exposes comprehensive SymPy functionality including:

• Symbolic Expressions: symbols, sympify, simplify, expand, factor
• Calculus: diff, integrate, limit, series, Derivative, Integral
• Solving: solve, solveset, nsolve, diophantine, dsolve
• Matrices: Matrix operations, inverse, determinant, eigenvalues
• Trigonometry: sin, cos, tan, and their inverses
• Special Functions: gamma, zeta, erf, fibonacci, etc.
• Number Theory: isprime, prime, gcd, lcm, totient
• Sets: FiniteSet, Interval, Union, Intersection
• Logic: And, Or, Not, Xor
• Printing: LaTeX, MathML, Python code

Development

git clone https://github.com/daedalus/mcp-sympy.git
cd mcp-sympy
pip install -e ".[test]"

# run tests
pytest

# format
ruff format src/ tests/

# lint
ruff check src/ tests/

# type check
mypy src/

License

MIT License - see LICENSE file.