pdf-to-markdown-cli 0.5.2

CLI utility to convert PDFs and supported document formats to Markdown/JSON/HTML with Marker API.

PDF to markdown CLI

Command-line utility for converting PDFs and other supported documents into Markdown, JSON, or HTML using the Marker API.

Why use this tool

• Converts single files or entire directories
• Automatically splits large PDFs into chunks and merges results
• Persists request state locally so interrupted runs can recover
• Rewrites and copies extracted images into deterministic output folders
• Supports OCR/LLM tuning flags from the Marker API

Supported formats

Input

• PDF (.pdf)
• Word (.doc, .docx, .odt)
• PowerPoint (.ppt, .pptx, .odp)
• Spreadsheets (.xls, .xlsx, .ods)
• EPUB/HTML (.epub, .html)
• Images (.png, .jpg, .jpeg, .webp, .gif, .tiff)

Output

• Markdown (.md, default)
• JSON (.json)
• HTML (.html)

Installation

pip install pdf-to-markdown-cli

From source:

git clone https://github.com/SokolskyNikita/pdf-to-markdown-cli.git
cd pdf-to-markdown-cli
pip install -e .

Quick start

export MARKER_PDF_KEY="your_api_key"
pdf-to-md ./examples/equations.pdf

Process a directory:

pdf-to-md ./docs

Use JSON or HTML output:

pdf-to-md ./examples/equations.pdf --json
pdf-to-md ./examples/equations.pdf --html

CLI options

• input: input file or directory path
• --json: output JSON instead of Markdown
• --html: output HTML instead of Markdown
• --langs: comma-separated OCR languages (default: English)
• --llm: enable LLM-enhanced processing
• --strip: redo OCR
• --noimg: disable image extraction
• --force: force OCR on all pages
• --pages: include page delimiters
• --max: enable all OCR enhancement flags (--llm --strip --force)
• -mp, --max-pages: process only the first N pages
• --no-chunk: disable PDF chunking
• -cs, --chunk-size: PDF pages per chunk (default: 25)
• -o, --output-dir: absolute output directory path
• -v, --verbose: debug logging
• --version: show installed package version

Development

Run tests:

python -m unittest discover -s tests -v

For contributions or questions, open a GitHub issue.