Skip to main content
Glama

PlotMCP Server

PlotMCP is a powerful Model Context Protocol (MCP) server designed to enable LLMs to generate high-quality SVG charts from structured data. It leverages fastmcp for the server infrastructure and matplotlib for consistent, precise chart rendering.

Key Features

  • Pure SVG Rendering: Generates static SVG format with no external JavaScript dependencies. Safe, portable, and easy to embed.

  • Multiple Plot Types: Supports Line, Scatter, Bar, Area, Histogram, Box, Heatmap, Contour, and Pie charts.

  • Flexible Configuration: Full control over titles, dimensions, margins, and axis properties (linear, log, and symlog scales).

  • Output Management: When --output-dir is configured, automatically saves generated charts and returns a specially formatted response that clients can parse to display the image:

    ```local_image
    /path/to/chart.svg
    ```

    This format allows clients to easily detect and render the generated images.

  • Deterministic Output: Ensures identical inputs produce bit-identical SVG outputs.

Related MCP server: mcp-plots

Installation

Requires Python >= 3.11 and uv installed.

Local Installation (Development)

git clone <repository-url>
cd plot-mcp
uv sync

Install as a Global Tool

uv tool install .

Running the Server

Running from Source

uv run plot-mcp --output-dir ./plots

Running Remotely via GitHub (using uvx)

You can run the server directly from the GitHub repository without manual cloning:

uvx --from git+https://github.com/Nexo-Agent/plot-mcp plot-mcp --output-dir ./plots

Note: Replace the URL with the actual repository location.

CLI Configuration

The server supports the following command-line options:

  • --output-dir PATH: Directory where generated SVG files will be saved. When set, tools return the file path instead of the raw SVG content.

  • --transport [stdio|sse|streamable-http]: The communication protocol (default: stdio).

  • --port INTEGER: The port for SSE or HTTP transport (default: 8000).

Output Format

The server supports two output modes depending on whether --output-dir is configured:

Without --output-dir (Default)

Tools return a PlotOutput object containing the raw SVG content:

{
  "svg": "<svg>...</svg>",
  "width": 800,
  "height": 400,
  "viewBox": "0 0 800 400"
}

With --output-dir (Recommended)

Tools save the SVG to a file and return a specially formatted string:

```local_image
/absolute/path/to/chart.svg
```

This format is designed to be easily parsed by clients. When your client receives a response containing this pattern, it should:

  1. Detect the ```local_image marker

  2. Extract the file path

  3. Load and display the image from that path

This approach keeps the response lightweight and allows clients to handle image rendering efficiently.

See examples/local_image_format.py for a complete demonstration of how this format works.

Available Tools

The LLM can invoke the following tools:

  1. plot_line: Render continuous 2D lines.

  2. plot_scatter: Render discrete 2D points.

  3. plot_bar: Render categorical bar charts.

  4. plot_area: Render filled area under a curve.

  5. plot_histogram: Render 1D histograms.

  6. plot_box: Render box plots from raw values.

  7. plot_heatmap: Render 2D matrix as a color grid.

  8. plot_contour: Render 2D contour lines.

  9. plot_pie: Render circular pie and donut charts.

Chart Configuration

All tools accept a shared config object to customize the visual output:

{
  "title": "My Chart",
  "width": 800,
  "height": 400,
  "margin": { "top": 40, "right": 20, "bottom": 40, "left": 50 },
  "x_axis": { "label": "X Axis", "scale": "linear" },
  "y_axis": { "label": "Y Axis", "scale": "log" }
}

License

MIT

Available Tools

9 tools
plot_areaC

Render filled area under a curve. Simple flat parameters - no nested objects!

Example: { "x": [1, 2, 3, 4], "y": [1, 4, 9, 16], "title": "Area Plot", "fill_color": "steelblue" }

Inputs: params (required). Returns: result

plot_barC

Render categorical bar chart. Simple flat parameters - no nested objects!

Example: { "categories": ["A", "B", "C"], "values": [10, 20, 15], "title": "Bar Chart", "orientation": "vertical" }

Inputs: params (required). Returns: result

plot_boxC

Render box plot from raw values. Simple flat parameters - no nested objects!

Example: { "groups": [ {"name": "Group A", "values": [1, 2, 3, 4, 5]}, {"name": "Group B", "values": [2, 3, 4, 5, 6]} ], "title": "Box Plot" }

Inputs: params (required). Returns: result

plot_contourB

Render 2D contour lines from grid data. Simple flat parameters - no nested objects!

Example: { "x": [1, 2, 3], "y": [1, 2, 3], "z": [[1, 2, 3], [4, 5, 6], [7, 8, 9]], "title": "Contour Plot" }

Inputs: params (required). Returns: result

plot_heatmapB

Render 2D matrix as color grid. Simple flat parameters - no nested objects!

Example: { "matrix": [[1, 2, 3], [4, 5, 6], [7, 8, 9]], "x_labels": ["X1", "X2", "X3"], "y_labels": ["Y1", "Y2", "Y3"], "title": "Heatmap" }

Inputs: params (required). Returns: result

plot_histogramC

Render 1D histogram. Simple flat parameters - no nested objects!

Example: { "values": [1, 2, 2, 3, 3, 3, 4, 4, 5], "bins": 5, "title": "Histogram" }

Inputs: params (required). Returns: result

plot_lineC

Render one or more continuous 2D lines. Simple flat parameters - no nested objects!

Example: { "series": [{"name": "Line 1", "x": [1, 2, 3], "y": [1, 4, 9]}], "title": "My Plot", "width": 800, "height": 400 }

Inputs: params (required). Returns: result

plot_pieC

Render circular pie chart. Simple flat parameters - no nested objects!

Example: { "labels": ["A", "B", "C"], "values": [30, 50, 20], "title": "Pie Chart" }

Inputs: params (required). Returns: result

plot_scatterC

Render discrete 2D points. Simple flat parameters - no nested objects!

Example: { "x": [1, 2, 3, 4], "y": [1, 4, 9, 16], "title": "Scatter Plot", "color": "steelblue" }

Inputs: params (required). Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.6/5.0

    Scored across 9 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose corresponding to a specific plot type (area, bar, box, contour, heatmap, histogram, line, pie, scatter). The names directly indicate the visualization type, making it impossible to confuse one tool for another. There is no functional overlap between tools as each creates a fundamentally different chart format.

    Naming Consistency5/5

    All tools follow the exact same 'plot_' prefix + chart_type naming pattern (plot_area, plot_bar, plot_box, etc.). This consistent verb_noun structure (where 'plot' is the verb and the chart type is the noun) creates perfect predictability. Every tool name uses the same snake_case convention without any deviations.

    Tool Count5/5

    Nine tools is ideal for a plotting server, covering the most common visualization types without being overwhelming. Each tool represents a distinct, essential chart type that earns its place in the set. The count aligns perfectly with the domain scope of providing core plotting functionality.

    Completeness5/5

    The tool set provides comprehensive coverage of standard plot types including area, bar, box, contour, heatmap, histogram, line, pie, and scatter plots. This covers the fundamental visualization needs for data analysis and presentation. There are no obvious gaps in the plotting domain - agents can create any common chart type needed for data visualization tasks.

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      C
      maintenance
      An MCP (Model Context Protocol) server that enables LLMs to generate ECharts visualizations by accepting chart type, data and parameters and returning cloud image URLs of the generated charts.
      12 npm
      87
      Apache 2.0
    • A
      license
      A
      quality
      D
      maintenance
      A MCP server for data visualization. It exposes tools to render charts (line, bar, pie, scatter, heatmap, etc.) from data and returns plots as either image/text/mermaid diagram.
      2
      4
      MIT
    • F
      license
      B
      quality
      D
      maintenance
      An MCP server that provides data visualization and machine learning tools, featuring automated intent-based pipeline routing for data cleaning and model training. It enables LLMs to process CSV or JSON data to generate visual charts, perform regressions, or execute clustering analysis.
      16
      -
    • A
      license
      A
      quality
      D
      maintenance
      An MCP server that enables AI assistants to create, update, and publish Datawrapper charts through natural language. It provides tools for data synchronization, visual configuration, and retrieving chart images or editor links.
      8
      MIT