SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillspdf-toolkit

pdf-toolkit

Office & productivity

View repository

Updated 10-16 09:17

About this skill

PDF Toolkit

Manipulate and read PDF files using open-source Python libraries: pypdf (BSD) for page operations, pdfplumber (MIT) for text/table extraction, and OCRmyPDF (MPL-2.0) for OCR.

When to Use

  • Extract text or tables from a PDF.
  • Merge several PDFs into one, or split / extract page ranges.
  • Rotate pages or add a watermark/stamp.
  • OCR a scanned (image-only) PDF to make it searchable.

For converting a PDF to clean Markdown for analysis, prefer the markitdown-converter skill. Use this skill for structural PDF operations.

Setup

pip install pypdf pdfplumber
# OCR (optional) also needs the system tools:
pip install ocrmypdf            # plus Tesseract + Ghostscript installed

How to Use

A bundled CLI (scripts/pdf_tools.py) covers the common operations:

# merge
python scripts/pdf_tools.py merge a.pdf b.pdf c.pdf -o combined.pdf

# split into one file per page, OR extract specific ranges
python scripts/pdf_tools.py split in.pdf -o pages_dir/
python scripts/pdf_tools.py split in.pdf --ranges 1-3 8 10-12 -o excerpt.pdf

# rotate (90/180/270)
python scripts/pdf_tools.py rotate in.pdf --deg 90 -o rotated.pdf

# extract text (stdout or -o file)
python scripts/pdf_tools.py text in.pdf -o out.txt

# watermark every page with a 1-page stamp PDF
python scripts/pdf_tools.py watermark in.pdf --stamp mark.pdf -o stamped.pdf

Tables

For table extraction, use pdfplumber directly — it exposes per-page tables:

import pdfplumber
with pdfplumber.open("report.pdf") as pdf:
    for page in pdf.pages:
        for table in page.extract_tables():
            print(table)   # list of rows

OCR a scanned PDF

ocrmypdf input_scanned.pdf output_searchable.pdf

This adds a searchable text layer (requires Tesseract + Ghostscript installed on the system).

Notes & Limits

  • pypdf handles structure (pages, rotation, merge, overlays) but does not re-flow or re-render content.
  • pdfplumber reads the embedded text layer; for image-only PDFs run OCR first.
  • Encrypted PDFs may need a password (PdfReader(path).decrypt(pw)).

Credits

Original skill by Rinu (l3ad3r1) in collaboration with Claude (Anthropic). Powered by pypdf (BSD-3-Clause), pdfplumber (MIT), and OCRmyPDF (MPL-2.0). All credit for the underlying capabilities belongs to those projects' authors.