matplotlib-viz-mcp.md

Matplotlib Visualization MCP

MCP server

from"xlisp"

xlisp maintains this community Python server for creating Matplotlib charts from MCP calls. Each tool saves a 300-DPI PNG in the machine's temporary directory and calls Matplotlib's local display window; it does not register an embedded chat interface.

Not for: Headless inline chart interaction. The server opens a local Matplotlib window and returns a saved-file path rather than a ui:// MCP App.

Source and setup verified on 2026-11-05

Visualization

matplotlib3DscatterPython

View on GitHub

Quick facts

Maintainer: xlisp (community)
Transport: stdio
Output: Local Matplotlib window plus 300-DPI PNG in the temp directory
Tools: 7 chart tools
UI: No inline MCP App

What you can do

create_relationship_graph
create_scatter_plot and create_classification_plot
create_3d_plot
create_histogram and create_line_plot
create_heatmap

Requirements

Python 3 with a virtual environment
mcp[cli], httpx, matplotlib, numpy, pandas and networkx
A desktop session able to open Matplotlib windows

setup.md

# 1. Clone the repository

$git clone https://github.com/xlisp/visualization-mcp-server.git

# 2. Create the environment

$python3 -m venv .venv --upgrade-deps

# 3. Install dependencies

$.venv/bin/pip install 'mcp[cli]' httpx matplotlib numpy pandas networkx

# Claude Desktop → claude_desktop_config.json

{
  "mcpServers": {
    "visualization": {
      "command": "/absolute/path/visualization-mcp-server/.venv/bin/python",
      "args": ["/absolute/path/visualization-mcp-server/visualization_server.py"]
    }
  }
}

Use .venv\Scripts\python.exe on Windows and replace both absolute paths.

Always confirm the latest setup instructions with the maintainer before installing.