Invoice & Receipt Extractor
Turns a folder of invoices and receipts (PDF or image) into a clean, categorized expense spreadsheet — vendor, dates, tax, and totals extracted, anomalies flagged, source files renamed. For bookkeepers, accountants, and anyone prepping expenses for the books.
Invoice & Receipt Extractor
Before You Begin
- A folder of invoice or receipt files (PDF, PNG, or JPG).
- The
xlsxskill available for building the workbook. - Optional: your GL codes or expense categories (a simple list or file). If omitted, a standard category set is used.
- Optional: the reporting period you're covering, so out-of-period dates get flagged.
- Write access to the folder if you want the source files renamed in place.
How It Works
- Direct the Skill to a folder. Example prompts:
- "Extract all the invoices in /receipts/q1 and build an expense spreadsheet."
- "Organize my receipts and categorize them using my GL codes."
- "Get the data out of these PDFs and flag anything that needs review."
- The skill inventories the folder and notes any file that can't be read.
- The skill extracts every field from each document — vendor, invoice number, invoice and due dates, line items, subtotal, tax (with type), total, currency, and payment method — using vision for image and scanned receipts.
- The skill cleans vendor names — strips legal suffixes, reference codes, and URLs, and fixes capitalization.
- The skill categorizes each expense against your categories, marking anything uncertain as
REVIEWwith a reason. - The skill flags anomalies — amounts over $5,000, duplicate invoice numbers, out-of-period dates, and missing or illegible totals.
- The skill builds the workbook and renames the processed source files.
What You Get
expenses.xlsx(3 sheets):- Sheet 1 — All expenses: date, vendor, invoice #, category, subtotal, tax, total, currency, payment method, source file, notes.
- Sheet 2 — Summary by category: count and totals (excl. and incl. tax) per category, with a grand total.
- Sheet 3 — Review required: every flagged item with the reason.
- Renamed source files in the format
YYYY-MM-DD_VendorName_$Total.pdf(unprocessable files prefixedUNREAD_and left in place). - A console summary: "Processed X files. Y need review. Total spend: $Z."
Troubleshooting
Related use cases
Turns how an accounting task was actually performed into an evidence-grade, sourced record — every action, every piece of context, and the rationale behind each, with gaps flagged rather than invented. For controllers and teams documenting close processes for audit or handoff.
Maintains a fixed asset register and runs depreciation end-to-end — processes additions and disposals, calculates depreciation with proration, and outputs a full schedule plus a ready-to-post journal entry. For controllers and accountants managing fixed assets each period.
Turns a folder of invoices and receipts (PDF or image) into a clean, categorized expense spreadsheet — vendor, dates, tax, and totals extracted, anomalies flagged, source files renamed. For bookkeepers, accountants, and anyone prepping expenses for the books.