by Anthropic
A comprehensive PDF manipulation toolkit from Anthropic that enables Claude to extract text and tables, create new PDFs, merge and split documents, fill forms, add watermarks, handle encryption, and perform OCR on scanned documents. The skill packages Python libraries (pypdf, pdfplumber, reportlab, pytesseract) and command-line utilities (pdftotext, qpdf, pdftk) into a modular Agent Skill that Claude loads on-demand whenever PDF-related tasks are detected. It supports progressive disclosure through three documentation layers: the core SKILL.md for common operations, FORMS.md for advanced form-filling workflows (including both fillable and non-fillable PDF handling), and REFERENCE.md for specialized libraries like pypdfium2 and pdf-lib. As one of Anthropic's pre-built Agent Skills, it works across Claude.ai, the Claude API, and Claude Code without requiring manual invocation.
Installation
Key Features
- ✓Extracts text with layout preservation and tables as pandas DataFrames using pdfplumber, supporting both structured and scanned PDF documents
- ✓Creates new PDF documents from scratch with styled text, shapes, and multi-page report layouts using ReportLab's Canvas and Platypus engines
- ✓Merges, splits, rotates, and manipulates existing PDFs with pypdf, including metadata extraction and page-level operations
- ✓Fills both interactive (fillable) and non-interactive PDF forms through a structured workflow with automated field detection, bounding box validation, and coordinate-based annotation
- ✓Performs OCR on scanned PDFs using pytesseract and pdf2image to make non-searchable documents fully text-extractable
- ✓Applies watermarks, password encryption/decryption, and image extraction using a combination of Python libraries and command-line tools (pdftotext, qpdf, pdftk)
Use Cases
- →Extracting structured data tables from financial reports, invoices, or research papers and converting them to spreadsheet-ready DataFrames
- →Automatically filling government, legal, or business PDF forms by detecting form fields and mapping values programmatically
- →Merging multiple PDF documents into a single file or splitting a large PDF into individual page files for distribution
- →Converting scanned paper documents into searchable, text-extractable PDFs through OCR processing
- →Generating formatted PDF reports, certificates, or multi-page documents from data with custom styling and layouts