pdfminer.six

Project Overview

pdfminer.six is a community maintained fork of the original PDFMiner for extracting information from PDF documents, with a focus on getting and analyzing text data. It pulls page text directly from the PDF's internal structure and can also recover exact position, font, and color of each string; the codebase is modular so every component can be swapped for custom pipelines.

Mine text, positions, fonts, and colors out of PDFs in pure Python.

Key Features

• Text extraction: parses the PDF internals to pull text straight off pages
• Layout analysis: automatically analyzes layout, reporting position, font, and color
• Multiple outputs: export content as text, HTML, or hOCR
• Image extraction: pulls embedded JPG, PNG, TIFF, and JBIG2 images
• CJK and vertical writing: supports Chinese, Japanese, Korean, and vertical text
• Encrypted documents: decrypts PDFs protected with RC4 or AES
• Form extraction: reads AcroForm interactive form data
• Outline extraction: recovers the table of contents and tagged contents

Target Audience

• Document automation and RPA developers
• Engineering teams digitizing archives and analytics data
• Internationalization projects handling multilingual PDFs
• Developers building their own PDF parsing components

Quick Start

1. Install

pip install pdfminer.six

2. Extract text from the command line

pdf2txt.py example.pdf

3. Call the high-level API in code

from pdfminer.high_level import extract_text

text = extract_text('example.pdf')
print(text)

4. Install the extra dependency when images are needed

pip install 'pdfminer.six[image]'

Core Advantages

• Pure Python: no system-level PDF library needed, deployment stays simple
• Deep detail: beyond raw text, positions, fonts, and colors are available
• Modular design: parsing, interpretation, and rendering components are independently replaceable

System Requirements

Notes

• Scans do not work: image-only scanned PDFs have no text layer and need OCR alongside
• Complex layouts need tuning: tables and multi-column pages may require custom layout analysis