# Ratnaditya-J/csvglow [Health: Active]

**Category:** 📊 Data Visualization  
**Repository:** https://github.com/Ratnaditya-J/csvglow  
**GitHub Stars:** 13  
**npm Downloads (last month):** 46  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/ratnaditya-j-csvglow

## Description
Generate beautiful self-contained HTML dashboards from CSV/Excel files with interactive ECharts visualizations, dark gradient theme, and sortable data tables.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "csvglow": {
    "command": "npx",
    "args": ["-y","csvglow"]
  }
}
```

## Documentation

## What Ratnaditya-J/csvglow MCP server does

Ratnaditya-J/csvglow MCP server turns a local tabular file into an interactive HTML dashboard. The generated output is a single file that includes its visualizations and interface, so it can be opened in a browser and used offline without a separate application server or CDN.

The package accepts CSV and TSV files, with delimiter detection for those formats, as well as XLS and XLSX workbooks. XLSX processing currently uses the first sheet. The dashboard includes a sortable and filterable table showing the first 1,000 rows, along with visual analysis based on the detected column types.

## How it works

The MCP mode runs locally and lets an MCP-compatible assistant request a dashboard from a file path. The underlying command can also be used directly from a terminal. When processing a file, csvglow identifies numeric, categorical, date/time, and identifier columns, then chooses relevant views automatically.

Numeric fields can receive histograms with descriptive statistics, including mean, median, standard deviation, quartiles, and outlier counts. Categorical fields can be represented with bar charts. Date columns can produce time-series charts with area fills. The generator also creates categorical-versus-numeric cross-analysis, a numeric correlation heatmap, and scatter plots for pairs whose absolute correlation exceeds 0.7.

The output includes multi-column findings that compare metrics and may point out contradictions, efficiency gaps, or high- and low-performing records. Large inputs may use smart sampling, including files with 100,000 or more rows. Each chart has a control for copying it to the clipboard.

## Setup and configuration

Install the package from PyPI with `pip install csvglow`, or run it through npx without a separate package installation. The direct command accepts an input file, an optional output path, and a flag to suppress automatic browser opening. For example, `csvglow data.csv -o dashboard.html` writes to a chosen location, while `csvglow data.csv --no-open` avoids opening the result automatically.

To expose Ratnaditya-J/csvglow MCP server through npx, configure the client with command `npx` and arguments `-y`, `csvglow`, and `--mcp`. The README documents configuration locations for Cursor at `.cursor/mcp.json` in a project and Windsurf at `~/.windsurf/mcp.json`. An existing pip installation can instead use the `csvglow` executable with the `--mcp` argument.

## Tools and capabilities

The documented MCP use case is dashboard generation from a file path. Resulting dashboards can contain:

- Automatically selected histograms, bar charts, time-series views, cross-analysis charts, heatmaps, and scatter plots
- Narrative findings based on relationships across multiple columns
- A sortable and filterable table limited to the first 1,000 rows
- Clipboard copying for individual charts
- A self-contained HTML result suitable for offline viewing

## Limitations and notes

Multi-sheet Excel processing is not supported yet; only the first XLSX sheet is used. Multi-file inputs, join keys, light themes, custom color palettes, and PDF export are listed as roadmap items rather than current capabilities. The repository describes the software as MIT-licensed. Ratnaditya-J/csvglow MCP server requires access to the local file path supplied for processing, and the provided material does not document remote file retrieval or database connections.

_Full upstream README: https://allmcps.com/mcp/ratnaditya-j-csvglow/readme_