Last verified
PDF Skill
Editor's PickExtract text, fill forms, and summarize PDF documents inside Claude Code sessions
When it triggers
Use when extracting text or filling PDF forms
"Extract a 60-page contract, summarize the key clauses, and flag anything unusual. One session."
What is the PDF Skill?
The PDF Skill is an Anthropic-built Claude Code skill that adds structured PDF handling to any Claude Code session. It extracts text from text-based and OCR-scanned PDFs, fills AcroForm fields, and produces page-range summaries, covering the most common productivity PDF workflows without requiring external tools.
How does it work?
The skill bundles scripts and reference libraries that Claude Code loads when the trigger phrase is detected. When you ask Claude to work with a PDF, the skill reads the file using its bundled parser, extracts content as structured text with page markers, and makes that content available to Claude’s context window. For form filling, it maps field names in the AcroForm dictionary to the values you provide and writes a new output file.
For workflows spanning multiple file formats, the PDF skill pairs with the XLSX skill for spreadsheet data and the DOCX skill for Word documents, covering the most common enterprise formats in one session.
When should you use it?
Use the PDF skill when you need to extract information from reports, contracts, or scanned forms; summarize long documents; or automate PDF form filling at scale. It works best when the PDFs live locally or are accessible via a file path the session can reach.
Frequently asked questions
Can the PDF skill extract text from scanned (image-based) PDFs?
Yes. The skill runs OCR on image-based pages using a bundled OCR library. Accuracy depends on scan quality: clean, high-resolution scans produce reliable output, while low-resolution or heavily compressed PDFs may have character errors that Claude will flag in the output.
How does the PDF skill handle multi-page documents?
All pages are processed by default, with extracted text concatenated in page order. You can specify a page range (e.g., pages 3-7) in your request to limit extraction to one section, which is practical for large PDFs where only one chapter is relevant.
Can the PDF skill write back to a PDF to fill form fields?
Yes, for PDFs with AcroForm interactive fields. The skill detects field names, accepts values from Claude's response or your explicit instructions, and writes a new PDF with the fields populated. The original file is not modified; a separate output file is created.