This server wraps SymPy's symbolic math engine so Claude can manipulate equations, solve systems, differentiate and integrate symbolically, work with matrices, and handle number theory operations. You get direct access to functions like simplify, solve, diff, integrate, and matrix operations through the MCP interface. Reach for this when you need Claude to actually compute derivatives, factor polynomials, solve equations symbolically, or generate LaTeX from mathematical expressions rather than just talking about math. It exposes over a dozen tool categories including calculus, linear algebra, trigonometry, special functions, and logic operations. The Python API is also available if you want to use these tools directly in your code without going through MCP.
MCP server that exposes SymPy's symbolic mathematics functionality
pip install mcp-sympy
mcp-name: io.github.daedalus/mcp-sympy
Configure in your MCP client:
{
"mcpServers": {
"mcp-sympy": {
"command": "mcp-sympy",
"env": {}
}
}
}
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"
The MCP server exposes comprehensive SymPy functionality including:
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/
MIT License - see LICENSE file.