Directmedia MCP

📖 Installation Guide — quick start, manual setup, and troubleshooting

FastMCP 3.1.0+ server for accessing Directmedia Publishing "Digitale Bibliothek" - TEXT EXTRACTION WORKING!

Quick Start

git clone https://github.com/sandraschi/directmedia-mcp
cd directmedia-mcp
just

This opens an interactive dashboard showing all available commands. Run just bootstrap to install dependencies, then just serve or just dev to start.

Manual Setup

If you don't have just installed:

Related MCP server: alexandria-mcp

Overview

The Directmedia Publishing "Digitale Bibliothek" was a pioneering German electronic book collection from the 1990s, containing extensive German literature and world literature. This MCP server provides programmatic access to these classic digital books.

BREAKTHROUGH: Text Extraction Working!

MISSION ACCOMPLISHED: We successfully reversed the Directmedia TEXT.DKI format!

• Discovery: TEXT.DKI files contain structured binary records, not compressed data

• Decompressor: Working Python implementation extracts readable German text

• Access: 101 volumes of 1990s literature now programmatically accessible

• Preservation: Digital cultural heritage unlocked for modern use

What was thought to be "compression" was actually a structured record format with 2-byte length headers!

Collection Status

• 101 volumes discovered (DB002-DB161, DBSK01-DBSK05, DBSO01-DBSO28)

• ~14GB total content across all volumes

• Proprietary binary format from 1990s German publishing

• Latin-1 encoding with special characters for German texts

Legal Requirement

You must legally purchase the Directmedia CD-ROMs to use this tool. See Legal Notice section below.

Sample Volumes

Collection Analysis

101 volumes discovered with ~50GB total content:

• DB002-DB061: Main literature collection (philosophy, literature, history)

• DBSK01-DBSK05: Schnellkurs (crash courses)

• DBSO01-DBSO28: Sonderausgaben (special editions)

File Format Structure

Each volume uses a proprietary binary format:

Core Files (Data/):

• TEXT.DKI: Main text database (structured binary records)

• TREE.DK*: Navigation tree (table of contents)

• INDEX.*: Multiple search indices (HTX, PLX, SHX, SWX, TTX, WLX)

• LINKS.*: Hyperlinks and cross-references

• SIGEL.DAT: Abbreviations/signatures registry

Media Files:

• IMAGES/: BMP illustrations and diagrams

• WAVS/: Audio files (readings, lectures)

• TABLES/: Specialized content tables

Quick Start

Prerequisites

• Python 3.11+

• Access to Directmedia "Digitale Bibliothek" collection

• FastMCP 3.1.0+

Installation

Prerequisites

• uv installed (RECOMMENDED)

• Python 3.12+

Quick Start

Run immediately via uvx:

uvx directmedia-mcp

Claude Desktop Integration

Add to your claude_desktop_config.json:

"mcpServers": {
  "directmedia-mcp": {
    "command": "uv",
    "args": ["--directory", "D:/Dev/repos/directmedia-mcp", "run", "directmedia-mcp"]
  }
}

Basic Usage

from directmedia_mcp import DirectmediaLibrary

# Initialize library
lib = DirectmediaLibrary(r"L:\Multimedia Files\Written Word\Digitale Bibliothek")

# List all volumes
volumes = lib.list_volumes()
print(f"Found {len(volumes)} volumes")

# Search for content
results = lib.search_text("Nietzsche", "DB002")  # Philosophy volume

# Extract text
content = lib.get_text_content("DB002", 0, 1000)

MCP Server Usage

# Start MCP server
python -m directmedia_mcp.server --library-path "L:\Multimedia Files\Written Word\Digitale Bibliothek"

# Or run directly
directmedia-mcp --library-path "L:\Multimedia Files\Written Word\Digitale Bibliothek"

MCP Tools

Library Management

• set_library_path(path) - Configure library location

• list_volumes() - List all available volumes

• get_volume_info(volume_id) - Get volume metadata

Content Access

• search_text(query, volume_id, limit) - Search across volumes

• get_text_content(volume_id, start_pos, length) - Extract text

• get_navigation_tree(volume_id) - Get table of contents

EPUB Conversion NEW

• convert_volume_to_epub_file(volume_id, output_dir) - Convert single volume to EPUB

• batch_convert_to_epub(output_dir, volume_ids) - Convert multiple volumes to EPUB

Analysis

• analyze_volume_structure(volume_id) - File format analysis

Volume Overview

Technical Details

Binary Format Analysis

TEXT.DKI Structure:

• Header: 256 bytes with section offset table

• Content: Structured binary records (not compressed!)

• Each record: 2-byte length + 1-byte type + text content

TREE.DK Structure:*

• DKA: Navigation tree with entry counts and offsets

• DKI: Tree structure data

INDEX Files:

• HTX: Hypertext index for navigation

• PLX: Plaintext index for full-text search

• SHX/SWX: Specialized search indices

• TTX: Title index

• WLX: Word list index

Known Limitations

1. Proprietary Format: No official documentation available

2. Advanced Features: Some INDEX and TREE.DK* structures still being analyzed

3. Encoding: Primarily Latin-1 with some UTF-8 elements

4. Media Content: Images and audio files not yet processed

Recent Achievements

• TEXT.DKI Decompression: Successfully reversed structured binary record format

• Text Extraction: Working decompressor extracts readable German text

• EPUB Conversion: Convert volumes to modern e-book format

• MCP Integration: Full programmatic access via FastMCP server

• Volume Management: Complete 101-volume library access

• TREE.DKI Navigation: Table of contents successfully parsed

EPUB Conversion Feature

Convert extracted Directmedia text content into modern EPUB format for e-book readers!

What It Does

• Extracts readable text from Directmedia .DKI files

• Formats content with proper HTML structure and CSS styling

• Creates valid EPUB 3.0 files compatible with all e-book readers

• Preserves German text encoding and special characters

• Adds metadata including title, author, and volume information

EPUB Features

• Proper Structure: Mimetype, container.xml, OPF package, navigation

• German Typography: Optimized for German text with proper quotes and spacing

• Responsive Design: CSS styling that works on all devices

• Table of Contents: Navigation structure for easy browsing

• Metadata: Complete Dublin Core metadata for library management

Usage Examples

Convert single volume:

# Via MCP tool
convert_volume_to_epub_file("DB002", "./epub_output")

Batch convert multiple volumes:

# Via MCP tool
batch_convert_to_epub("./epub_library", ["DB002", "DB003", "DB004"])

Output Example

epub_output/
 Goethe - Faust.epub          # Volume DB004
 Heine - Buch der Lieder.epub # Volume DB007
 ... (more volumes)

EPUB Reader Compatibility

• Calibre (recommended for library management)

• Apple Books (iOS/macOS)

• Google Play Books

• Kindle (via conversion)

• Adobe Digital Editions

• All major e-book readers

Future Enhancements

• Complete INDEX file parsing for full-text search

• TREE.DK* advanced structure decoding

• Cross-volume search optimization

• Image extraction and processing

• Audio file handling

Contributing

This is a research project to preserve and provide access to classic digital literature. Contributions welcome for:

• Binary format analysis

• Decompression algorithms

• Search optimization

• Documentation improvements

Legal Notice & Copyright

Important: Legal Use Required

This software tool is designed to work with legally purchased copies of Directmedia Publishing's "Digitale Bibliothek" CD-ROM collection. You must own legitimate copies of the CD-ROMs to use this tool legally.

Where to Purchase

Directmedia Publishing still operates and offers their complete collection:

• Official Website: https://www.directmedia-publishing.de/

• Product: "Digitale Bibliothek" (Complete 101-volume collection)

• Format: Available as digital downloads and physical media

• Languages: German literature and philosophy collections

Copyright Notice

• Copyright: Directmedia Publishing GmbH

• Content: All text, images, and multimedia content remain copyrighted

• Usage: Personal, educational, and research use permitted with legal copies

• Redistribution: Not permitted without explicit permission

Disclaimer

This tool is provided for educational and research purposes to access legally obtained digital content. The authors are not responsible for misuse of this software. Ensure you comply with all applicable copyright laws in your jurisdiction.

Pirated or illegally obtained content is not supported and may violate copyright law.

🛡️ Industrial Quality Stack

This project adheres to SOTA 14.1 industrial standards for high-fidelity agentic orchestration:

• Python (Core): Ruff for linting and formatting. Zero-tolerance for print statements in core handlers (T201).

• Webapp (UI): Biome for sub-millisecond linting. Strict noConsoleLog enforcement.

• Protocol Compliance: Hardened stdout/stderr isolation to ensure crash-resistant JSON-RPC communication.

• Automation: Justfile recipes for all fleet operations (just lint, just fix, just dev).

• Security: Automated audits via bandit and safety.

License

MIT License - see LICENSE file for details.

Acknowledgments

• Directmedia Publishing for pioneering electronic literature in the 1990s

• The German digital humanities community

• FastMCP framework for MCP implementation

Webapp Dashboard

This MCP server includes a free, premium web interface for monitoring and control. By default, the web dashboard runs on port 10826. (Assigned ports: 10826 (Web dashboard frontend), 10827 (Web dashboard backend (API)))

To start the webapp:

1. Navigate to the webapp (or web, frontend) directory.

2. Run start.bat (Windows) or ./start.ps1 (PowerShell).

3. Open http://localhost:10826 in your browser.