Does this skill require an API key to function?

While simple manual conversions can be performed without one, high-accuracy VLM parsing requires a MinerU API key (MINERU_API_KEY) to handle complex tables and figures.

What is the difference between VLM and Pipeline mode?

VLM mode uses a vision-language model for 90%+ accuracy on complex layouts but is slower, while Pipeline mode uses traditional parsing for faster processing of simpler documents.

Can it accurately parse academic papers with multiple columns?

Yes, the skill is specifically designed for complex layouts. When using MinerU in VLM mode, it maintains structural integrity for multi-column academic papers.

How does it handle non-PDF formats like audio?

The skill includes fallback logic and integration workflows for audio, recommending external transcription services like Otter.ai or Rev.com for specialized media.

Can I convert specific pages of a large document?

Yes, the skill supports page selection parameters (e.g., 'pages 1-10') to minimize costs and focus on relevant sections of a document.

Document Conversion & Parsing

Name: Document Conversion & Parsing
Author: linxule

bylinxule

•

데이터 과학 및 ML

Converts complex PDFs, academic papers, and tables into structured formats using intelligent tool selection and VLM-based parsing.

This skill provides robust document processing capabilities within Claude Code, specializing in transforming academic papers and complex PDFs into research-ready data. It intelligently toggles between high-accuracy Vision-Language Model (VLM) parsing via MinerU for intricate layouts—including merged tables, formulas, and figures—and lightweight manual methods for simpler documents. By automating the extraction of structured content from multi-column layouts, it ensures qualitative researchers and developers can quickly ingest literature and data into their AI workflows without losing structural integrity.

주요 기능

01High-accuracy VLM-based parsing for complex academic and technical layouts

02Multi-column support and automated OCR error cleanup

03Precision extraction of tables, mathematical formulas, and figures

042 GitHub stars

05Intelligent tool selection based on document complexity and API availability

06Batch processing support for managing multiple document queues

사용 사례

01Converting structured PDF reports into clean markdown for AI processing

02Digitizing legacy field notes and multi-column documents with visual elements

03Importing dense academic papers with tables for qualitative research analysis

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add linxule/interpretive-orchestration document-conversion

For use in Claude.ai and ChatGPT

Download Skill