docx2img

A pure Python .docx → image rendering engine. No Microsoft Office or LibreOffice required.

Features

• Paragraphs, runs, fonts, colors, bold/italic/underline/strikethrough
• Full style inheritance and theme resolution
• Tables with merged cells, nested tables, and borders
• Inline images and floating images with text wrap
• Multi-section documents, landscape pages, multi-column layout
• Headers, footers, and page numbers
• Ordered / unordered lists with numbering
• Word-style pagination with widow/orphan control and cross-page tables
• CJK line breaking, punctuation hanging, justification, and document-grid layout
• Justified text, tab stops, text boxes
• OMML math (fractions, super/subscripts, radicals, accents, bars, limits, equation arrays)
• Footnotes and endnotes with continuation pages
• DrawingML and legacy VML shapes, text boxes, flowchart nodes, and arrows

Installation

pip install docx2img

Usage

CLI

docx2img input.docx output.png --dpi 150

Python API

from docx2img import Config, convert, convert_to_images

# A multi-page document is saved as output_1.png, output_2.png, ...
convert("input.docx", "output.png", dpi=150)

# Per-page images
pages = convert_to_images("input.docx", Config(dpi=150))
for i, img in enumerate(pages):
    img.save(f"page_{i + 1}.png")

Development

pip install -e ".[dev]"
python -m pytest tests/ -v

See docs/technical_design.md for architecture details.

License

MIT