PDF

The PDF Claude skill does every common PDF job from a chat: merge, split, extract text and tables, OCR scanned files, fill forms, add watermarks, encrypt. Drop the file in. Get the result back.

The Problem

PDFs show up everywhere. Doing anything to them is somehow always harder than you remember.

You need to send the accountant twelve receipts as a single file. You google "merge pdf", land on a site plastered with ads, watch your scanned bank statement upload to a server you've never heard of, and get a watermarked output you can't actually use. You try Acrobat, twenty bucks a month for one job, or you download a free desktop tool that wants to install three other things. Half an hour later you've got one PDF and a vague sense of having been mugged.

Or you've got a vendor PDF with a pricing table you need in a spreadsheet. You copy. You paste. The columns are wrong, the line breaks are weird, half the numbers landed in the wrong row. You start typing it by hand.

This skill fixes that.

What this Claude skill does

• Merges PDFs into one file: combine 2 or 200 documents in the page order you want.
• Splits a PDF apart: every page as its own file, or specific ranges like 1-5 and 6-10.
• Extracts text and tables: pull the words out cleanly, or pull a table straight into a spreadsheet.
• OCRs scanned PDFs: turn an image-only PDF into searchable text, page by page.
• Fills forms, encrypts, rotates, watermarks: complete PDF forms, add or remove passwords, fix sideways scans, overlay a watermark on every page.
• Creates new PDFs from scratch: reports, letters, multi-page documents with headings, tables, page numbers, even proper subscripts and superscripts.

Why it's amazing!

• Stop installing yet another sketchy desktop PDF tool.
• Combine, extract, or repair a PDF in seconds instead of half an hour.
• Get clean tables out of vendor invoices without retyping a single cell.
• Never upload a sensitive document to a random "PDF tools online" site again.

Who it's for

Anyone who deals with PDFs. Accountants pulling tables out of vendor invoices, founders combining diligence docs, lawyers reading scanned contracts, ops teams archiving form submissions, and indie hackers building any product that touches a PDF file. Anyone who's ever lost a morning to merging twelve files.

How it works

Ask Claude something like:

• "Merge these 12 receipts into receipts-Q3.pdf."
• "Split contract.pdf into one file per page."
• "Extract every table from this vendor invoice and put them in an Excel file."
• "I have a scanned 30-page contract. OCR it so I can search for 'termination'."
• "Build a 2-page client report: title page, three sections, page numbers, our logo in the header."
• "Take watermark.pdf and stamp it on every page of proposal.pdf."

You'll get back something like:

Merged: receipts-Q3.pdf
  - 12 source files combined, 47 pages total
  - Bookmarks preserved, output size 4.2 MB

Split: contract-pages/
  - 30 single-page PDFs (page_1.pdf … page_30.pdf)
  - All metadata copied through

Extracted: vendor-invoice-tables.xlsx
  - 3 tables found across 5 pages
  - 187 rows, 6 columns (Item, SKU, Qty, Unit Price, Tax, Total)

OCR'd: contract-searchable.pdf
  - 30 pages processed, text layer added behind images
  - "termination" found on pages 7, 14, 22

Watermarked: proposal-final.pdf
  - "DRAFT" overlay applied to every page (47 pages)
  - Original layout untouched

Built: client-report.pdf
  - 2 pages, US Letter, header with logo
  - 3 sections, page numbers in footer

Everything runs locally. Your scanned bank statements, contracts, and tax returns never leave your machine.

FAQ

Does it cost anything to run?

It needs a Claude subscription but outside of that there are no external premium APIs or subscriptions needed. It uses open-source Python libraries (pypdf, pdfplumber, reportlab) and command-line tools (qpdf, poppler) you install once. No subscriptions, no per-file fees.

How accurate is the text and table extraction?

For PDFs with a real text layer, extraction is near-perfect. For scanned (image-only) PDFs, OCR runs through Tesseract, strong on clean scans, weaker on faded receipts or handwriting. Tables come out clean when the source has consistent borders or alignment. Messy multi-column layouts may need a second pass.

Does my PDF ever get uploaded to a random PDF site?

No. The PDF file itself is processed locally by open-source tools on your machine, never uploaded to a third-party PDF service. Note that Claude Code does send your conversation to Anthropic to work, so if you ask Claude to read, extract, or analyze a PDF's contents, that extracted text is part of the request like any other chat. But the file is never handed to some ad-covered "PDF tools online" site.

Can it edit text inside an existing PDF?

Not freely. PDFs aren't built for inline text editing the way Word documents are. You can fill form fields, overlay watermarks, add or rotate pages, or rebuild the PDF from scratch. For real text edits, extract the content, edit it with the docx skill, then re-export.

Does it work on password-protected PDFs?

Yes, if you know the password. It can also add or remove passwords on PDFs you own.

How many PDFs can it handle at once?

Hundreds. Merging 200 files or splitting a 500-page report works the same as one or two — the tools stream pages rather than loading everything into memory.

Does it preserve bookmarks, links, and form fields?

Yes, on merge and split operations. Bookmarks copy through, internal links survive when the target page survives, and form fields stay live unless you flatten them on purpose.

Can it convert a PDF to Word or Excel?

Tables come out as Excel-ready files. For full PDF-to-Word conversion, extract the text here and rebuild the document with the docx skill. The round-trip is cleaner than any direct converter.

Safe To Install

Please read before installing. Any skill you copy online is someone else's code running inside your Claude, and a lot of what's out there should give you pause. Most authors who give skills away don't give a second thought to quality or security. Most are vibe-coded slop pushed out untested, and some are the occasional bad actor hiding an exploit in plain sight. (Read about OpenClaws ClawHub security nightmare). Every skill we publish is vetted by a programmer with 20+ years of professional experience and checked line by line for anything that touches your files without reason. With Stim-Pack Studios Claude Skills, you get peace of mind knowing you're installing safe, vetted code, not a stranger's guess.

Add it to your Claude workflow

If you handle PDFs at all (invoices, contracts, scanned documents, client reports), this one earns its slot. It replaces the sketchy upload sites, the Acrobat subscription, and the half-installed desktop tools you forgot you have.

Get the PDF Claude skill and handle every PDF job from a chat.