Does it require a GPU to function?

No, a GPU is not required. While GPU acceleration helps with complex OCR tasks, the skill includes several high-performance CPU backends for standard text-based PDFs.

Can I process multiple PDFs at once?

Yes, it supports batch extraction for entire directories with a built-in resume feature that skips already processed files to save time.

What happens if one extraction method fails?

The skill features an automatic fallback system that tries up to 9 different backends (such as pdfplumber, markitdown, or pdfminer) in a prioritized order until it succeeds.

How does it handle scanned documents?

It uses vision-based backends like 'marker' and 'docling' which perform OCR and layout analysis, specifically optimized for scanned documents when a GPU is available.

Is the output optimized for Claude and other LLMs?

Yes, the tool generates clean markdown files, which provide the structural context and formatting necessary for LLMs to process document data accurately.

PDF Data Extractor

Name: PDF Data Extractor
Author: ahundt

byahundt

•

Data Science & ML

Extracts structured text and tables from PDF documents using multiple intelligent backends to create LLM-ready markdown.

The PDF Data Extractor skill empowers Claude to parse and convert complex PDF documents into clean, structured markdown. It leverages a multi-backend approach—including Microsoft’s MarkItDown, IBM’s Docling, and vision-based OCR—to ensure high-quality extraction from digital text, complex tables, and even scanned images. Whether you are processing batch invoices, academic papers, or insurance policies, this skill automatically selects the best available engine (with GPU acceleration support) to handle layout analysis and data recovery, making it an essential tool for RAG workflows and automated document analysis.

Key Features

01Automatic GPU detection and acceleration for vision-based OCR tasks

026 GitHub stars

03Batch processing capabilities with resume functionality for large directories

04Intelligent layout analysis for complex tables, columns, and academic papers

05Multi-backend fallback system with 9 specialized extraction engines

06Standardized markdown output optimized for LLM processing and RAG

Use Cases

01Automating data extraction from multi-page insurance policies and legal documents

02Extracting text and mathematical equations from academic PDFs for research summaries

03Converting a directory of scanned invoices into structured markdown for data analysis

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ahundt/autorun pdf-extractor

For use in Claude.ai and ChatGPT