Extract structured JSON data from Hebrew government forms using Tesseract OCR.
npx skills add https://github.com/skills-il/localization --skill hebrew-ocr-forms
Please help me install this Agent Skill. Skill: hebrew-ocr-forms Source: https://github.com/skills-il/localization/tree/main/hebrew-ocr-forms Command: npx skills add https://github.com/skills-il/localization --skill hebrew-ocr-forms
Hebrew government forms scanned for OCR produce unstructured text with mixed Hebrew/English, making manual data extraction slow and error-prone. This skill provides automated detection and extraction of form fields from common Israeli government forms (e.g., Tabu, Tofes 106, Tofes 857) and outputs structured data for downstream workflows, including validation of Israeli ID numbers and RTL text handling.
Preprocess a scanned Hebrew form and run the extractor to obtain structured JSON data.
High-intent search queries and answers about installing and using this skill.
Extract structured data from scanned Hebrew government forms by applying image preprocessing and Tesseract OCR with Hebrew and English languages. The process handles RTL text, detects form types like Tabu or Tofes 106, and outputs validated JSON fields.
Tesseract OCR handles mixed Hebrew and English text in RTL scanned images by configuring the heb+eng language parameters. This setup correctly processes right-to-left layouts and mixed content commonly found in Israeli government documents.
Automate data extraction from Israeli Tax Authority forms by using form-type detectors to identify Tofes 106 and 857 documents. The system then applies specific field extractors to capture tax year and employer details, outputting structured JSON data.
This OCR data extraction approach validates Israeli ID numbers automatically. After extracting text using Tesseract, the system applies ID validation logic to ensure the structured JSON output contains accurate and verified identification data.
You need Python image processing libraries like Pillow, opencv-python, and numpy to preprocess scanned forms before OCR. These dependencies handle image cleaning and formatting to improve Hebrew text recognition accuracy.
Limitations when processing Tabu land registry documents with OCR include potential accuracy issues with heavily degraded scans and mixed Hebrew/English layouts. While preprocessing and RTL handling mitigate common errors, raw OCR text output is available for manual verification.
ZurybrExtract and interpret code from images using OCR and image preprocessing.
alibabaExtract text, tables, and metadata from PDF documents into JSON.
legoutExtract text and structured data from images and scanned PDFs using Tesseract OCR.
meytalp-devExtract structured fields from scanned Hebrew government forms using OCR.
