Extracts structured financial data from invoice images and PDFs using advanced dual-verification OCR processing.
The reading-invoice skill automates the tedious task of manual bookkeeping by converting invoice images and PDF files into clean, structured data. It features a unique dual-verification workflow that processes images through two independent contexts to ensure high accuracy for critical financial fields like total amounts and registration numbers. Specifically optimized for Japanese tax compliance (shinkoku), it handles Japanese era dates, consumption tax rates, and invoice registration numbers (T+13), making it an essential tool for developers building automated accounting or e-Tax preparation pipelines.
主な機能
01Extraction of T+13 invoice registration numbers and tax breakdowns
02Seamless PDF text extraction with automatic fallback to image OCR
03323 GitHub stars
04Automatic conversion of Japanese era names (Reiwa/Heisei) to Western dates
05Dual-verification system for high-precision financial data extraction
06Batch processing support for multiple files and directory-wide scanning
ユースケース
01Automating data entry for Japanese tax returns (Kakutei Shinkoku)
02Digitizing paper receipts into structured accounting software formats
03Verifying invoice compliance with the Japanese qualified invoice system
