invoice-processor_skill

Overview

This skill automates extraction of structured data from invoice PDFs and images and produces formatted Excel spreadsheets. It combines OCR and AI vision to detect invoices, extract fields, and generate a clean Excel report ready for accounting or analysis. Use it to batch-process invoice files and convert them to machine-readable spreadsheets.

How this skill works

The tool scans a specified folder for invoice files (PDF, JPG, PNG), converts PDFs to images, and runs AI vision/OCR to extract key invoice fields. It processes files concurrently (configurable limits), outputs a JSON results file with success/error statuses, and then converts that JSON into a formatted Excel workbook. Environment checks ensure the API key and required Python packages are present.

When to use it

• You receive batches of invoices to digitize or archive
• You need invoices converted from PDF/image to Excel for accounting
• You want automated extraction of invoice fields (date, number, amounts, parties)
• You need a reproducible workflow for recurring invoice processing
• You must handle invoices in directories with non-ASCII names and require path-safety

Best practices

• Place all invoice files in a dedicated input folder and use the script input path option for custom locations
• Create a .env file with your API key (GLM_API_KEY) before running scripts
• Always run the included environment check script to verify dependencies (aiohttp, PyMuPDF, openpyxl)
• Use full relative script paths when the project directory name contains non-ASCII characters (avoid short paths)
• Keep file sizes and resolutions within limits (≤5MB, ≤6000x6000 px) and respect API concurrency limits

Example use cases

• Monthly accounting: batch convert last month’s paper invoices to Excel for ledger import
• AP automation: extract invoice fields automatically and feed them into an ERP or bookkeeping system
• Audit preparation: create a single formatted Excel of all invoice metadata for reviewers
• Data extraction for analytics: pull line items and amounts to build spend reports