
Math Mcp
Brings GPU-accelerated scientific computing to Claude through SymPy symbolic algebra, NumPy numerical operations, and FFT/optimization routines. The core math server exposes 14 tools including symbolic_solve for equation solving, symbolic_diff for calculus, matrix operations with automatic CUDA detection, and numerical integration. It's part of a larger four-server system (math, quantum, molecular, neural) that shares data via URIs, so you can solve differential equations symbolically then pass results to quantum simulations or molecular dynamics. Runs via uvx without installation or pip install scicomp-math-mcp. GPU support cuts matrix multiply times from 2 seconds to 35ms on 4096x4096 operations. Async task support handles long-running computations without blocking.
- GitHub stars
- 2
- tools
- 3
- transport
- STDIO
- last updated
- Jan 5, 2026
Tools
3 toolsPublic tool metadata for what this MCP can expose to an agent.
symbolic_solveSolve equations symbolically.
Solve equations symbolically.
No parameter schema in public metadata yet.
symbolic_diffCompute derivatives.
Compute derivatives.
No parameter schema in public metadata yet.
matrix_multiplyGPU-accelerated matrix operations.
GPU-accelerated matrix operations.
No parameter schema in public metadata yet.
Sponsors
Readme
Math-Physics-ML MCP System
GPU-accelerated Model Context Protocol servers for computational mathematics, physics simulations, and machine learning.
📚 Documentation
Details
- Repository
- github.com/andylbrummer/math-mcp
- Website
- andylbrummer.github.io/math-mcp/api/math-mcp
- Category
- Developer Tools
- Package
scicomp-math-mcp- Transport
- STDIO
- Registry
- MCP registry, active