AI Invoice Data Extraction with Python & GPT

Every month we get 100 to 200 supplier invoices as PDFs, and we type the key details into a spreadsheet by hand. We need a Python script that does this with AI. What the script should do: - Read all PDF invoices from a folder - Use the OpenAI API (GPT-4o or similar) to extract: - Supplier name - Invoice number - Invoice date - Due date - Subtotal, tax, and total amount - Currency - Save the results to one CSV file, one row per invoice Technical details: - Invoices come from about 20 different suppliers, so layouts vary - Most PDFs have selectable text. A few are scanned images. - Use structured JSON output from the model so the fields are consistent - Dates in one format (YYYY-MM-DD) and amounts as plain numbers - If a field can't be found, leave it blank and flag the row for review - One bad file should not stop the whole run - We already have an OpenAI API key Deliverables: - Python source code + requirements.txt - README with setup and how to run it - Test run on 10 sample invoices we'll provide, with the output CSV Not included: Accounting software integration, a web interface, or email inbox automation.