Fermat MCP

This project provides a FastMCP server for mathematical computations, including numerical and symbolic calculations, as well as plotting.

Modules

1. mpl_mcp - Matplotlib Integration

Feature	Description
plot_barchart	Plots bar charts of given data values
plot_scatter	Creates scatter plots from data points
plot_chart	Plots line, scatter, or bar charts
plot_stem	Creates stem plots for discrete data
plot_stack	Generates stacked area/bar charts
eqn_chart	Plots mathematical equations

2. numpy_mcp - NumPy Integration

Category	Operations
Basic Math	add, sub, mul, div, power, abs, exp, log, sqrt
Trigonometric	sin, cos, tan
Statistics	mean, median, std, var, min, max, argmin, argmax, percentile
Linear Algebra	dot, matmul, inv, det, eig, solve, svd
Matrix Operations	create, zeros, ones, full, arange, linspace
Array Manipulation	reshape, flatten, concatenate, transpose, stack

3. sympy_mcp - SymPy Integration

Category	Operations
Algebra	simplify, expand, factor, collect
Calculus	diff, integrate, limit, series
Equations	solve, solveset, linsolve, nonlinsolve
Matrix Operations	create, det, inv, rref, eigenvals

Setup

Requirements

• Python 3.12 or higher (To install Python3.12 follow Python Download)

• uv (To install uv follow uv Installation)

Clone the repository

git clone https://github.com/abhiphile/fermat-mcp

Visual Studio Code, Windsurf

You can find the mcp.json file in the MCP: Open User Configuration or MCP: Open Workspace Configuration

Add the following to your mcp.json:

{
  "mcpServers": {
    "fmcp": {
      "command": "bash",
      "args": ["MCP_SERVER_ABSOLUTE_PATH/setup.sh"],
      "description": "fmcp server is for mathematical computations, including numerical and symbolic calculations, as well as plotting."
    }
  }
}

Claude (Anthropic)

If you're using Claude or the Anthropic MCP client, add this working MCP configuration to your mcp.json (update the directory path to your local clone):

{
  "mcpServers": {
    "fmcp": {
      "command": "uv",
      "args": [
        "--directory",
        "/home/ty/Repositories/fermat-mcp",
        "run",
        "server.py"
      ]
    }
  }
}

Gemini CLI

• Open your Gemini settings JSON located in ~/.gemini/settings.json where ~ is your home directory.

• Add the following to your settings.json:

{
  "mcpServers": {
    "fmcp": {
      "command": "bash",
      "args": ["MCP_SERVER_ABSOLUTE_PATH/setup.sh"],
      "description": "fmcp server is for mathematical computations, including numerical and symbolic calculations, as well as plotting."
    }
  }
}

Installing via Smithery

To install Fermat MCP for local usage automatically via Smithery:

npx -y @smithery/cli install @abhiphile/fermat-mcp --client gemini

Example Usage

• Using Gemini CLI

╭──────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│  > Can you use fmcp server and using numpy method find the eigen values of this 8*8 matrix,                  |
│    2 1 3 1 1 8 4 2                                                                                           |
│    6 6 0 7 1 4 6 1                                                                                           │
│    9 2 1 8 7 9 9 0                                                                                           │
│    2 5 6 6 9 8 0 1                                                                                           │
│    1 3 6 2 3 8 8 1                                                                                           │
│    9 4 2 2 1 2 2 9                                                                                           │
│    8 6 4 4 2 0 2 8                                                                                           │
│    0 0 0 6 6 7 5 6                                                                                           │
╰──────────────────────────────────────────────────────────────────────────────────────────────────────────────╯

 ╭─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
 │ ✔  numpy_mcp_numerical_operation (fmcp MCP Server) {"a":[[2,1,3,1,1,8,4,2],[6,6,0,7,1,4,6,1],[9,2,1,8,7,9,9,0],[2,5,6,6,9,8,0,1],[1,3,… │
 │                                                                                                                                         │
 │    {"eigenvalues":["32.077244457548815+0j","-11.531090644775198+0j","-6.6653982146786195+0j","0.6715984762411508+3.37024850             │
 │    10270413j","0.6715984762411508-3.3702485010270413j","4.541270555490195+2.776364664923869j","4.541270555490195-2.77636466             │
 │    4923869j","3.6935063384423428+0j"],"eigenvectors":[["-0.23263835483680192+0j","-0.2264723575289234+0j","-0.4308391916391             │
 │    0195+0j","-0.012346573390129022+0.17748655663058255j","-0.012346573390129022-0.17748655663058255j","-0.21221572277027187             │
 │    +0.3524396218277479j","-0.21221572277027187-0.3524396218277479j","0.3451499664861578+0j"],["-0.31955742545335186+0j","-0             │
 │    .2569860493445581+0j","0.05691886770041556+0j","-0.35591013681869693-0.2242364092694275j","-0.35591013681869693+0.224236             │
 │    4092694275j","0.1932161673963751-0.39527849111641133j","0.1932161673963751+0.39527849111641133j","-0.7979681696063214+0j             │
 │    "],["-0.46626263247473404+0j","-0.4684914620112376+0j","0.5469400556350749+0j","0.34325164099973565+0.06607019711949293j             │
 │    ","0.34325164099973565-0.06607019711949293j","0.21312270185159682+0.28822307710358636j","0.21312270185159682-0.288223077             │
 │    10358636j","0.42707422750984786+0j"],["-0.41589316441674523+0j","0.2291771012892302+0j","0.09410792992600435+0j","0.6375             │
 │    92441360358+0j","0.637592441360358+-0j","0.46446646137729414+0j","0.46446646137729414+-0j","0.08171661775583623+0j"],["-             │
 │    0.35812884189789035+0j","-0.26551071423139044+0j","-0.649979374400915+0j","-0.2999153430497845+0.20110182336747695j","-0             │
 │    .2999153430497845-0.20110182336747695j","0.04424367733961273-0.020400357434930304j","0.04424367733961273+0.0204003574349             │
 │    30304j","-0.037276545325188024+0j"],["-0.3042416233100104+0j","0.4027405071320744+0j","0.2343447346668298+0j","-0.119224             │
 │    87219398595+0.10317726668781271j","-0.11922487219398595-0.10317726668781271j","-0.3387759141362585+0.15790241128043445j"             │
 │    ,"-0.3387759141362585-0.15790241128043445j","0.11473144194899779+0j"],["-0.35389732755448017+0j","0.5359362131480812+0j"             │
 │    ,"0.13989902586074168+0j","-0.16008513226461316-0.26111873108374495j","-0.16008513226461316+0.26111873108374495j","0.055             │
 │    729749650881726-0.22748184206919145j","0.055729749650881726+0.22748184206919145j","-0.19838514696566653+0j"],["-0.327615             │
 │    5213691367+0j","-0.30122843153433887+0j","0.07858579556736099+0j","0.00042251379361638556-0.11670182367986544j","0.00042             │
 │    251379361638556+0.11670182367986544j","-0.18590228871569267-0.2479130526139182j","-0.18590228871569267+0.247913052613918             │
 │    2j","-0.03374593856285834+0j"]]}                                                                                                     │
 ╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
✦ Of course, I can do that. Here are the eigenvalues for the given matrix:

  32.077244457548815
  -11.531090644775198
  -6.6653982146786195
  0.6715984762411508 + 3.3702485010270413j
  0.6715984762411508 - 3.3702485010270413j
  4.541270555490195 + 2.776364664923869j
  4.541270555490195 - 2.776364664923869j
  3.6935063384423428

• Using Visual Studio Code, Windsurf Here the task was to plot both x**2 and sin(x) in the same plot.