Any2md

any2md — Universal Document to Markdown

Automatically routes to the best backend based on file format and content:

| Format | Backend | Why | |--------|---------|-----| | DOCX with math | /word2md | Correct OMML→LaTeX, no escaping bugs | | DOCX no math | markitdown | General-purpose conversion | | PDF | markitdown | Table-aware extraction | | PDF (high fidelity) | PyMuPDF | --pdf-backend pymupdf | | PPTX / XLSX / HTML / EPUB / etc | markitdown | Broad format support |

When to Use

• User mentions: "any2md", "/any2md", "convert to md", "转 markdown", "万能转 md"
• User has a document of unknown type and wants Markdown output
• User wants one command that handles any format

How to Use

PYTHONIOENCODING=utf-8 python ~/.claude/skills/any2md/any2md.py "" [""]

Arguments

| Arg | Required | Description | |-----|----------|-------------| | input | Yes | Input file path | | output | No | Output .md path (default: stdout) | | --math | No | auto (default) / yes / no — math detection | | --pdf-backend | No | markitdown (default) / pymupdf |

Examples

# Auto-detect format, print to stdout
PYTHONIOENCODING=utf-8 python ~/.claude/skills/any2md/any2md.py "document.docx"

# Convert and save
PYTHONIOENCODING=utf-8 python ~/.claude/skills/any2md/any2md.py "slides.pptx" "slides.md"

# Force math mode for DOCX
PYTHONIOENCODING=utf-8 python ~/.claude/skills/any2md/any2md.py "report.docx" "report.md" --math yes

# PDF with PyMuPDF (high fidelity, no tables)
PYTHONIOENCODING=utf-8 python ~/.claude/skills/any2md/any2md.py "paper.pdf" "paper.md" --pdf-backend pymupdf

Key Features

1. Auto math detection: Scans DOCX for oMath elements; routes to word2md if found
2. Math escaping fix: Post-processes markitdown output to fix \_ → _ inside $...$
3. Image extraction: DOCX with math → word2md extracts images to images/ folder
4. PDF backends: Choose between table-aware (markitdown) or high-fidelity text (PyMuPDF)

Dependencies

• markitdown — pip install markitdown
• python-docx — for word2md backend (auto-detected)
• PyMuPDF (optional) — for --pdf-backend pymupdf

Rules

1. Always use PYTHONIOENCODING=utf-8 on Windows
2. For DOCX with math, this skill auto-routes to word2md — no need to call /word2md separately
3. For pure PDF reading (no conversion needed), prefer /pdf-reader which has 3 extraction modes
4. For DOCX to HTML (not Markdown), use /word2html instead

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: phanghonghao
• Source: phanghonghao/THU-Awesome-Skills
• License: MIT

Install and usage instructions live in the source repository linked above.