powerpoint-analyzer-mcp by shiosalt - MCP Server

PowerPoint Analyzer MCP Server

An MCP server that enables search and extraction using PowerPoint structure and text formatting attributes.

Background

Most AI Agent searches that claim PowerPoint support typically ignore PowerPoint file structure and only extract text for searching. This tool enables outputting text written in bold and other structured information.

Features

Text formatting detection: Bold, italic, underline, strikethrough, highlighting, hyperlinks
Font analysis: Font sizes, colors, and styling information
Slide querying: Query slides with flexible filtering criteria
Table data extraction: Extract table data with formatting detection
Testing suite for formatting detection validation
Implementation using Python standard libraries (no external PowerPoint dependencies)
Direct XML parsing for processing
Built with FastMCP 2.0

Project Structure

powerpoint-analyzer/
├── main.py                     # Main FastMCP server entry point
├── powerpoint_mcp_server/      # Core server implementation
│   ├── server.py              # Main MCP server implementation
│   ├── config.py              # Configuration management
│   ├── core/                  # Core functionality
│   └── utils/                 # Utility modules
├── tests/                      # Test files
│   ├── test_powerpoint_fastmcp.py  # Main server tests
│   ├── test_formatting_detection.py # Formatting detection tests
│   └── ...                         # Other test files
├── scripts/                    # Utility scripts
│   ├── health_check.py        # Server health check
│   └── start_server.py        # Alternative server startup
├── requirements.txt            # Python dependencies
├── pytest.ini                 # Test configuration
└── README.md                   # Documentation

Installation

Clone the repository:

git clone <repository-url>
cd powerpoint-analyzer

Install dependencies:

pip install -r requirements.txt

Configure your AI agent (Claude Desktop, etc.) by adding the following to your configuration file:

Location of mcp_settings.json:

macOS: ~/Library/Application Support/Claude/mcp_settings.json
Windows: %APPDATA%\Claude\mcp_settings.json

{
  "mcpServers": {
    "powerpoint-analyzer-mcp": {
      "command": "python",
      "args": ["/path/to/your/powerpoint-analyzer/main.py"],
      "env": {
        "POWERPOINT_MCP_LOG_LEVEL": "INFO"
      }
    }
  }
}

Example with actual paths:

macOS/Linux:

{
  "mcpServers": {
    "powerpoint-analyzer-mcp": {
      "command": "python",
      "args": ["/Users/username/powerpoint-analyzer/main.py"],
      "env": {
        "POWERPOINT_MCP_LOG_LEVEL": "INFO"
      }
    }
  }
}

Windows:

{
  "mcpServers": {
    "powerpoint-analyzer-mcp": {
      "command": "python",
      "args": ["C:\\Users\\username\\powerpoint-analyzer\\main.py"],
      "env": {
        "POWERPOINT_MCP_LOG_LEVEL": "INFO"
      }
    }
  }
}

Technical Approach

This server processes PowerPoint files using the following approach:

Direct ZIP handling: .pptx files are processed as ZIP archives using Python's zipfile module
XML parsing: Internal PowerPoint XML structure is parsed using xml.etree.ElementTree with namespace support
Dual formatting detection: Supports both XML attribute and child element formats for text formatting properties
No external dependencies: Uses only Python standard library modules for PowerPoint processing
Processing: Extracts only the required information without loading entire presentations into memory
Caching: Caching system for performance on repeated operations

Text Formatting Detection

The server provides text formatting detection capabilities:

Supported Formatting Types

Bold text: Detects bold formatting in text elements
Italic text: Identifies italic styling across slides
Underlined text: Finds underlined text with underline styles
Strikethrough text: Detects strikethrough formatting
Highlighted text: Identifies highlighted/background colored text
Hyperlinks: Extracts hyperlink information and relationship IDs
Font properties: Analyzes font sizes, colors (RGB and scheme colors)

Technical Implementation

Dual detection method: Checks both XML attributes (b="1") and child elements (<a:b val="1"/>)
Namespace-aware parsing: Handling of Office Open XML namespaces
Validation: Test suite for detection across different PowerPoint versions
Debug capabilities: Debugging tools for troubleshooting formatting detection issues

Validation and Testing

Test suite: tests/test_formatting_detection.py validates formatting types
Debug tools: tests/debug_formatting_detection.py provides XML analysis
Validation: Tested with PowerPoint files containing mixed formatting

Usage

Running the Server

python main.py

Available Tools

This MCP server provides three core tools:

extract_formatted_text: Extract text with specific formatting types (bold, italic, underline, strikethrough, highlight, hyperlinks, font sizes, font colors)
query_slides: Query slides with flexible filtering criteria
extract_table_data: Extract table data with flexible selection and formatting detection

Development

Requirements

Python 3.8+
MCP (Model Context Protocol)
FastMCP 2.0
Standard Python libraries (zipfile, xml.etree.ElementTree)

Recent Updates

Version 2.0 - Advanced Text Formatting Detection

Fixed critical formatting detection bug: Bold, italic, underline, and strikethrough attributes now correctly detected
Dual detection support: Handles both XML attribute and child element formats
Comprehensive test suite: Added extensive tests for all formatting types
Debug tools: New debugging utilities for troubleshooting formatting issues
Enhanced MCP tools: New tools for formatted text extraction and analysis

Formatting Detection Validation

All formatting types have been thoroughly tested and validated:

✅ Bold text: 8 elements detected in test files
✅ Italic text: 6 elements detected in test files
✅ Underlined text: 6 elements detected in test files
✅ Strikethrough text: 6 elements detected in test files
✅ Highlighted text: 5 elements detected in test files
✅ Hyperlinks: 2 elements detected in test files
✅ Font colors: Multiple colors detected and analyzed

License

This project is licensed under the Apache License 2.0 - see the file for details.