Does it support scanned documents that are not born-digital?

Absolutely. By enabling OCR and selecting the Granite or Tesseract engines, you can convert non-digital, scanned PDFs into structured, machine-readable text.

Can this skill handle multi-column documents?

Yes, Docling Advanced automatically detects multi-column layouts and preserves the logical reading order of the text during the conversion process.

What is the benefit of using the Granite model with Docling?

Granite is a vision-language model that significantly improves extraction accuracy for scanned PDFs and complex visual layouts compared to standard OCR engines.

What formats can the extracted data be exported to?

The skill supports exporting document content into several structured formats including Markdown, HTML, and JSONL.

How can I optimize Docling for faster processing?

You can optimize performance by disabling unused features like table structure analysis or OCR in the Pipeline Options and implementing parallel processing using ThreadPoolExecutor.

Docling Advanced Document Extraction

Name: Docling Advanced Document Extraction
Author: orbruno

byorbruno

0•

数据科学与机器学习

Extracts structured data from complex PDFs, scanned documents, and multi-column layouts using the advanced Docling engine and Granite vision-language models.

Docling Advanced empowers Claude to handle the most challenging document processing tasks by leveraging sophisticated OCR and vision-language models like Granite. It specializes in converting scanned PDFs, complex tables, and multi-column layouts into clean, structured Markdown or JSONL while providing fine-grained control over extraction pipeline options. Whether you are processing high-volume research papers or low-quality scanned invoices, this skill optimizes both performance and accuracy for professional-grade document intelligence and RAG preparation.

主要功能

01Performance optimization via parallel processing and granular feature toggling

02Formula and mathematical content identification within technical and academic documents

03High-fidelity table extraction for complex, merged, or multi-page table structures

04Granite vision-language model support for superior OCR and layout understanding

05Automated multi-column layout detection and logical reading order preservation

060 GitHub stars

使用场景

01Converting complex academic research papers into structured Markdown for RAG systems

02Digitizing legacy scanned invoices and financial statements with high structural accuracy

03Extracting large-scale data from multi-column technical manuals and corporate reports

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add orbruno/docling-ccplugin docling-advanced

For use in Claude.ai and ChatGPT

Download Skill