MCP Table Parser

A Model Context Protocol (MCP) server for parsing and analyzing tabular data from clipboard. Transform raw data into AI-ready structured format instantly.

Features

• Auto-detect delimiters - CSV, TSV, pipe-separated data
• Smart type inference - Numbers, dates, strings automatically recognized
• Data summary - Row/column counts, null value analysis
• Sample preview - First 3 rows for immediate understanding
• Pure input/output - No file handling, clipboard-to-insight pipeline

Installation

npm install -g mcp-table-parser

Usage

1. Configure Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "table-parser": {
      "command": "mcp-table-parser"
    }
  }
}

2. Use with Claude

Simply paste any table data and ask for analysis:

Help me analyze this sales data:
Product,Quantity,Price,Date
Apple,120,5.5,2024-01-15
Banana,85,3.2,2024-01-16
Orange,95,4.8,2024-01-17

Claude will automatically:

• Parse the data structure
• Identify column types
• Provide data summary
• Enable instant analysis

Supported Data Types

• Numbers - Automatically detected and validated
• Dates - Multiple format support
• Strings - Text and categorical data
• Mixed data - Heterogeneous columns handled gracefully

Supported Delimiters

• Comma (,) - CSV files
• Tab (\t) - TSV files, Excel copy-paste
• Pipe (|) - Custom format data
• Auto-detection - Let the server figure it out

Example Outputs

Input

Name,Age,City,Join Date
Alice,28,New York,2023-01-15
Bob,32,San Francisco,2023-02-20
Charlie,25,Chicago,2023-03-10

Output

{
  "success": true,
  "delimiter": ",",
  "summary": {
    "rowCount": 3,
    "columnCount": 4,
    "columns": [
      {"name": "Name", "type": "string", "nullCount": 0},
      {"name": "Age", "type": "number", "nullCount": 0},
      {"name": "City", "type": "string", "nullCount": 0},
      {"name": "Join Date", "type": "date", "nullCount": 0}
    ],
    "sampleData": [...]
  },
  "data": [...]
}

API

Tools

parse_table_data

Pastes and analyzes tabular data.

Parameters:

• data_text (string, required): The raw table data
• delimiter (string, optional): Specify delimiter or use 'auto' for detection

Returns:

• Structured data with type inference and summary statistics

License

MIT License - see LICENSE file for details.

Contributing

Pull requests welcome! Please ensure:

• Code passes TypeScript compilation
• Tests cover new functionality
• Documentation is updated

Support

Issues and feature requests: GitHub Issues