OCR and extract invoice/receipt data → validate → output organized spreadsheet. Use when processing invoices, receipts, or expense documents.
Source: $ARGUMENTS (defaults: ./inbox/ → ./output/invoices.xlsx)
Deterministic pipeline: scan documents → extract fields → validate → write spreadsheet.
$SOURCE_DIR for *.pdf, *.png, *.jpg, *.jpegvendor (company name)date (normalize to YYYY-MM-DD)amount (numeric, preserve currency symbol)invoice_number (as printed; null if absent)category (auto-classify: travel, office, software, meals, other)invoice_number across documentsamount = 0 or negativedate more than 90 days in the past or any future datevendor, date, amount)$OUTPUT_FILE with columns:
File | Vendor | Date | Amount | Currency | Invoice# | Category | Flag
Date ascending[FLAG: reason] in Flag column)$SOURCE_DIR/
*.pdf | *.png | *.jpg | *.jpeg — invoice/receipt files
$OUTPUT_FILE.xlsx — populated spreadsheet, sorted by date
stdout — summary: N docs processed, N flagged, totals by currency
[UNREADABLE] flag and continueEdit PDFs with natural-language instructions using the nano-pdf CLI.