vision-ocr

Convert handwritten and scanned PDFs into markdown with OCR engines.

90|3|Updated Apr 21, 2026
One-click install
npx skills add https://github.com/OPTIMETA/PAIDEIA --skill vision-ocr
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: vision-ocr
Source: https://github.com/OPTIMETA/PAIDEIA/tree/main/plugins/paideia/skills/vision-ocr
Command: npx skills add https://github.com/OPTIMETA/PAIDEIA --skill vision-ocr

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill turns handwritten or scanned answer PDFs into accurate markdown transcription so grading and review can happen without manual retyping.

Core Features & Use Cases

  • Three OCR engines: Uses Claude native vision by default, with opt-in Ollama Qwen3-VL and pytesseract fallback paths for different privacy and fidelity needs.
  • Grade-ready output: Preserves original language, math notation, problem numbering, and ambiguous glyph markers while skipping crossed-out work.
  • Practical scenarios: Ideal for converting exam submissions, worksheet scans, and messy handwritten answers into markdown files for /grade.

Quick Start

Use the vision-ocr skill to transcribe the attached scanned answer PDF into markdown while preserving the original language, math notation, and problem numbering.

Frequently Asked Questions about vision-ocr

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

FAQPage Schema
How do I convert a scanned PDF with handwritten answers into markdown?

To convert a scanned PDF with handwritten answers into markdown, use an OCR skill that transcribes image-based documents while preserving original language and math notation. It processes exam submissions and worksheet scans into clean markdown text for grading workflows.

Does OCR transcription preserve math notation and problem numbering from worksheets?

OCR transcription preserves math notation, original language, and problem numbering from worksheets. The process skips crossed-out work and includes ambiguity marking for uncertain glyphs to ensure reliable conversion for grading workflows.

Can I use Ollama or Tesseract instead of native vision for PDF transcription?

You can use Ollama Qwen3-VL or pytesseract as fallback paths for PDF transcription instead of native vision. These opt-in engines accommodate different privacy and fidelity needs when converting scanned documents into markdown.

What is the best way to transcribe exam answer scans for automated grading?

The best way to transcribe exam answer scans for automated grading is using deterministic OCR that converts image-based PDFs into clean markdown. This preserves problem numbering and math notation, creating files ready for grading workflows.

How does handwriting OCR handle ambiguous characters or crossed-out work?

Handwriting OCR handles ambiguous characters by adding ambiguity markers to the markdown output and skips crossed-out work entirely. This ensures the final transcription reflects only the intended answers for accurate grading.

Do I need to install dependencies to run Tesseract or Ollama for scanned document conversion?

You do not need mandatory dependencies to run scanned document conversion since Claude native vision is the default. Ollama and Tesseract are optional opt-in fallback paths for users with specific privacy or fidelity requirements.