PDF to Markdown Converter

A powerful PDF to Markdown converter with GUI preview, powered by the marker-pdf library. Features GPU acceleration, side-by-side PDF and Markdown preview, and export to both Markdown and Word formats.

• 📄 High-Quality Conversion: Converts PDF documents to Markdown with excellent formatting preservation
• 🖼️ Image Extraction: Automatically extracts and saves embedded images from PDFs
• 👁️ Live Preview: Side-by-side PDF and Markdown preview with page navigation
• 🚀 GPU Acceleration: 3-10x faster conversion on NVIDIA GPUs (RTX series)
• 📝 Multiple Export Formats: Export to Markdown (.md) or Word (.docx)
• 🎯 User-Friendly GUI: Simple tkinter-based interface for easy operation

The application features three main panels:

• Left: PDF canvas preview
• Right: Markdown text preview
• Bottom: Page navigation controls

• Windows OS (other platforms may work with minor modifications)
• Python 3.8 or higher
• (Optional) NVIDIA GPU with CUDA support for faster processing

1. Clone the repository

   git clone https://github.com/ToGoRun/PDF_To_Markdown.git
   cd PDF_To_Markdown

2. Run the application (automatic setup)

   .\run.bat

   The script will automatically:

   • Create a virtual environment
   • Install all required dependencies
   • Launch the application

For NVIDIA GPU users, enable GPU acceleration for 3-10x faster conversion:

# Check if GPU is detected
.\check_gpu.bat

# Install CUDA-enabled PyTorch
.\setup_gpu.bat

For high-quality Word export, install Pandoc:

# Run the installation helper
.\install_pandoc.bat

# Or download manually from:
# https://github.com/jgm/pandoc/releases/latest

Note: If Pandoc is not installed, the application will automatically use python-docx as a fallback.

1. Launch the application

   .\run.bat

2. Convert a PDF

   • Click "选择PDF文件" to select your PDF file
   • Click "开始转换" to start conversion
   • Use navigation buttons to preview different pages
   • Export using "导出Markdown" or "导出Word"

3. View Results

   • Markdown files are saved with extracted images in {filename}_images/ folder
   • Word documents include all content and images

• Conversion Engine: marker-pdf - State-of-the-art PDF to Markdown conversion
• PDF Rendering: PyMuPDF (fitz) for fast PDF preview
• GUI Framework: tkinter (built-in Python GUI library)
• GPU Acceleration: PyTorch with CUDA support

On first run, the application downloads AI models (~1-2GB) to the ./models/ directory. These models are cached for future use, so subsequent conversions are much faster.

The application automatically detects available compute devices:

• ✅ CUDA GPU (NVIDIA)
• ✅ CPU fallback

GPU information is displayed in the status bar when a GPU is detected.

Core libraries:

• marker-pdf - PDF to Markdown conversion engine
• PyMuPDF - PDF rendering and preview
• Pillow - Image processing
• python-docx - Word document generation
• pypandoc - Enhanced Word export (requires Pandoc binary)

PyTorch is automatically installed by marker-pdf. For GPU acceleration, run setup_gpu.bat to install the CUDA version.

Problem: Application shows "CPU" instead of your GPU

Solution:

.\check_gpu.bat    # Check current status
.\setup_gpu.bat    # Install CUDA-enabled PyTorch

Problem: "Export failed, need to install Pandoc" error

Solution:

• Run .\install_pandoc.bat for guided installation
• Or download Pandoc manually from https://github.com/jgm/pandoc/releases/latest
• The application will use python-docx fallback if Pandoc is unavailable

Problem: First conversion takes a long time

Explanation: The first run downloads ~1-2GB of AI models. This is normal and only happens once. Subsequent conversions use cached models and are much faster.

This project is licensed under the MIT License - see the LICENSE file for details.

• marker-pdf - Excellent PDF to Markdown conversion library
• PyMuPDF - Fast PDF rendering
• Pandoc - Universal document converter

Contributions are welcome! Please feel free to submit a Pull Request.

For issues and questions, please use the GitHub Issues page.

Note: This tool is designed for Windows. For Linux/Mac users, you may need to modify the batch scripts (.bat) to shell scripts (.sh).