pdf-extraction

Extract PDF text with positional coordinates and offset data.

1|Updated Dec 9, 2025
One-click install
npx skills add https://github.com/BinkyTwin/DeepRead --skill pdf-extraction
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pdf-extraction
Source: https://github.com/BinkyTwin/DeepRead/tree/main/.claude/skills/pdf-extraction
Command: npx skills add https://github.com/BinkyTwin/DeepRead --skill pdf-extraction

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pymupdf, pdfplumber.

What problem does it solve?

This Skill extracts text from PDFs with exact positions for each element, enabling precise citations and downstream data extraction.

Core Features & Use Cases

  • Position-aware extraction: Get text blocks with normalized coordinates and start/end offsets.
  • Structured output: Returns pages_data with pageNumber, textContent, textItems, width, height.
  • Use Case: Ingest academic papers or reports to enable exact quotes and data extraction.

Quick Start

Run the Python-based extraction on a PDF (e.g., extract_text_with_positions('document.pdf')) to obtain pages_data and inspect the first page's text items.

Frequently Asked Questions about pdf-extraction

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract text from a PDF with precise location data for citations?

Text extraction with coordinates enables citing exact passages. This Skill extracts text blocks with normalized positions (0-1) and start/end offsets, returning structured output per page so you can pinpoint and quote specific content reliably.

Can I get both text content and coordinate positions from a PDF in one operation?

Yes. The Skill processes each PDF page and returns textItems containing textContent paired with positional data, width, height, and pageNumber, letting you map text to its exact location without separate passes.

What output structure does PDF text extraction with positions provide?

The Skill returns pages_data with pageNumber, textContent, textItems (including startOffset and endOffset), width, height, and hasText flags, giving you both raw text and machine-readable position metadata for downstream processing.

How do I handle PDFs with sparse text or image-heavy pages during extraction?

The Skill supports optional OCR on text-sparse pages and per-page image extraction, detecting pages with minimal text via the hasText field and applying alternative extraction methods where needed.

What's the difference between basic text extraction and position-aware extraction for academic papers?

Position-aware extraction preserves normalized coordinates and character offsets alongside content, enabling exact citation generation and structured data linking—essential for academic workflows where quoting and reference accuracy matter.

Does PDF extraction with coordinates work for ingesting reports and research documents at scale?

Yes. The structured output format and per-page processing support batch ingestion of academic papers and reports, with coordinates enabling reliable downstream citation, entity extraction, and data integration workflows.