legal-ocr

Community

Turn scanned legal PDFs into editable text.

Authorfbmoulin
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This Skill converts scanned Brazilian legal PDFs (sentenças, petições, acórdãos) into editable text with high accuracy, enabling quick digitization of archives, searchable databases, and automated workflows.

Core Features & Use Cases

  • High-precision OCR: Optimized for Portuguese legal language to maximize accuracy on court documents.
  • Multi-engine OCR: PaddleOCR as the primary engine, with EasyOCR as automatic fallback for difficult scans.
  • Advanced pre-processing: Deskew, noise removal, and CLAHE to improve readability before OCR.
  • Structure detection: Automatically identifies sections like relatório, fundamentação, dispositivo for structured outputs.
  • Quality validation: Confidence scoring and review flags to streamline human-in-the-loop checks.

Quick Start

Process a scanned PDF with the pipeline, for example: python pipeline_ocr.py sentenca.pdf

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: legal-ocr
Download link: https://github.com/fbmoulin/legal-ocr-skill/archive/main.zip#legal-ocr

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository