Csvglow

What is Csvglow?

Csvglow generates beautiful, interactive HTML dashboards from CSV or Excel files with a single command and zero configuration. It produces a self-contained HTML file that opens in the browser, featuring auto-detected charts, multi-column narrative insights, correlations, and a sortable data table. It also functions as an MCP (Model Context Protocol) tool for AI-compatible clients and as an OpenClaw skill.

How to use Csvglow?

Install via pip install csvglow or run directly with npx (npx csvglow data.csv). Use csvglow data.csv to generate a dashboard and open it in the browser. Additional options include -o for a custom output path and --no-open to prevent auto-opening. For MCP server mode, add a configuration entry to your client’s MCP config file (e.g., .cursor/mcp.json) with "command": "npx" and "args": ["-y", "csvglow", "--mcp"] (or "command": "csvglow" if installed via pip).

Key features of Csvglow

• Smart multi-column narrative analysis with contradiction detection
• Histograms for numeric columns with stats sidebar
• Bar charts for categorical columns
• Automatic categorical × numeric cross analysis with mean lines
• Time series line charts with area fill for date columns
• Correlation heatmap and scatter plots for highly correlated pairs

Use cases of Csvglow

• Instantly visualize any CSV or Excel file as a rich dashboard
• Use as an MCP tool so an AI assistant can generate dashboards from file paths
• Embed a self-contained, offline HTML report with charts and data table
• Quickly share interactive dashboards with stakeholders (no server needed)
• Analyze large datasets (100k+ rows) with smart sampling

FAQ from Csvglow

What file formats does Csvglow support?

It supports .csv, .tsv (auto-detected delimiter), .xls, and .xlsx (first sheet only; multi-sheet support is planned).

Can Csvglow handle large files?

Yes, for files with 100k+ rows it uses smart sampling to keep the output performant.

How do I use Csvglow as an MCP server?

Add a JSON entry to your MCP-compatible client’s config file (e.g., .cursor/mcp.json) with the command npx -y csvglow --mcp or csvglow --mcp if installed via pip. No extra installation is needed when using npx.

Does the output require an internet connection?

No, the output is a single self-contained HTML file that works offline with no server or CDN dependencies.

Is there any installation required for the MCP mode?

When using npx, no installation is needed. If you prefer to install first, run pip install csvglow and use csvglow --mcp.