Parse PDFs and scans into Markdown with preserved headings and tables.
npx skills add https://github.com/memect/kc --skill document-parsing-memect
Please help me install this Agent Skill. Skill: document-parsing Source: https://github.com/memect/kc/tree/main/template/skills/zh/document-parsing Command: npx skills add https://github.com/memect/kc --skill document-parsing-memect
Parse source documents into machine-readable text with maximum fidelity. Use when processing any document in Samples or Input for the first time, when parsed text quality is poor, or when tables and charts need special handling. Covers multi-level parser selection from simple text extraction to OCR and vision models. Also use when a verification rule fails due to parsing issues (garbled text, missing tables, mangled layouts) and the parser needs to be upgraded for that document type.
Feed a sample document to the parser and receive a clean Markdown rendition with preserved structure and tables.
High-intent search queries and answers about installing and using this skill.
To extract text and tables from a PDF into clean Markdown, the parser applies multi-level text extraction, OCR, and vision parsing to preserve headings, lists, and page boundaries. It evaluates document complexity to select the optimal parsing level for high-fidelity structured output.
The best way to parse scanned documents with complex layouts is using Level 3 local tools like MineRU alongside VLM-based OCR. This approach handles multi-page documents and intricate structures, outputting clean machine-readable text while maintaining structural fidelity.
Document text extraction results in garbled text and mangled layouts when simple Level 1 extraction fails on complex documents. The parser applies quality checks to detect these issues and automatically escalates parsing levels to VLM-based OCR or local tools for accurate extraction.
Yes, you can use VLM-based OCR for multi-page document processing. The parser supports Level 2 vision parsing to handle scans and complex layouts across multiple pages, producing clean Markdown with preserved page boundaries and structural elements.
You need to escalate parsing levels for source documents when quality checks detect garbled text, missing tables, or mangled layouts. The system automatically upgrades from Level 1 direct text extraction to Level 2 OCR or Level 3 local tools to ensure high-fidelity machine-readable text.
juliuss1907Extract text from PDFs, scanned documents, and images using OCR.
LuisElPotiExtract clean markdown or text content from web URLs via Tavily CLI.
joonlabExtract text from images and PDFs using the DeepSeek-OCR-8bit model.
niranjannavExtract text, tables, and metadata from PyMuPDF-compatible PDF documents.
