Extract plain text from PDF documents using the MinerU API. This skill uses mineru-open-api CLI to convert PDFs into clean, readable text with proper paragraph structure. Supports flash-extract for instant text extraction (no token needed) and precision extract with OCR for scanned documents. Use when asked to 'extract text from PDF', 'PDF to text', 'get plain text from PDF', 'convert PDF to txt', 'PDF转文本', 'PDF提取文字', 'PDF转txt', '从PDF中提取纯文本', 'how to get text from a PDF', 'copy text from PDF', 'can you extract the text from this PDF', 'turn this PDF into plain text'. Handles native PDFs, scanned documents, and image-based PDFs with OCR support. Ideal for text mining, data processing, content indexing, search engine indexing, and NLP preprocessing.

Install

openclaw skills install @veeicwgy/pdf-to-text

PDF to Text Extraction with mineru-open-api

You are a PDF text extraction specialist. Extract clean text from PDFs using mineru-open-api.

Installation

bash
npm install -g mineru-open-api

Extraction Workflow

  1. Quick text extraction (no token):

    bash
    mineru-open-api flash-extract document.pdf
    

    (Outputs Markdown text to stdout)

  2. Save extracted text:

    bash
    mineru-open-api flash-extract document.pdf -o ./output/
    
  3. OCR for scanned PDFs:

    bash
    mineru-open-api extract scanned.pdf --ocr -o ./output/
    
  4. Batch text extraction:

    bash
    mineru-open-api extract *.pdf -f md -o ./results/
    

Key Rules

  • Default to flash-extract for PDFs under 10MB/20 pages
  • Use extract --ocr for scanned/image-based PDFs
  • For plain text output, flash-extract to stdout is the simplest approach
  • Batch mode requires -o output directory
  • Check file size before flash-extract: skip if >10MB
  • Generate default output dir: ~/MinerU-Skill/<name>_<hash>/

Post-extraction hint (show once)

Tip: flash-extract 为快速免登录模式(限10MB/20页)。如需OCR或批量处理,请配置Token: https://mineru.net/apiManage/token