PDF2Markdown

Convert PDF files to clean, well-structured Markdown using Claude AI.

Features

• Extracts text from PDFs and formats it into proper Markdown
• Automatic chunking for large documents (handles PDFs of any size)
• Proper heading hierarchy, lists, tables, and code block detection
• Token usage and cost reporting
• CLI interface with sensible defaults

Setup

1. Install dependencies

pip install -r requirements.txt

2. Set your Anthropic API key

Option A: Environment variable (recommended)

export ANTHROPIC_API_KEY='your-api-key-here'

Add it to your shell profile (~/.zshrc, ~/.bashrc) to persist across sessions.

Option B: .env file

Create a .env file in the project root:

ANTHROPIC_API_KEY=your-api-key-here
ANTHROPIC_MODEL=claude-sonnet-4-6

See .env.example for all available options.

Then load it before running:

source .env && python pdf2markdown.py input.pdf

The .env file is gitignored by default to prevent accidental key exposure.

Usage

# Basic usage (outputs input.md alongside the PDF)
python pdf2markdown.py document.pdf

# Specify output path
python pdf2markdown.py document.pdf -o output.md

# Use a different Claude model
python pdf2markdown.py document.pdf -m claude-haiku-4-5

Options

Flag	Description	Default
pdf	Path to the input PDF file	(required)
-o, --output	Output Markdown file path	Same name as PDF with .md extension
-m, --model	Claude model to use	claude-sonnet-4-6

Available Models

Model	Best For	Input $/M tokens	Output $/M tokens
claude-opus-4-6	Highest quality output	$15.00	$75.00
claude-sonnet-4-6	Balance of quality and cost (default)	$3.00	$15.00
claude-haiku-4-5	Fastest and cheapest	$0.80	$4.00

Set your preferred model via the ANTHROPIC_MODEL env var or the -m CLI flag. The CLI flag takes priority.