The problem
Manhead Merchandise sells merch on tour, and every show ends with a settlement: a PDF from atVenu with the night's sales and taxes.
Those numbers have to reach NetSuite, the company's ERP, in the shape NetSuite imports.
What I built
Python and Tesseract OCR read each settlement PDF, multi-page ones included, and write NetSuite-ready CSV files, with tax split by venue and sales split by artist.
An OpenAI-assisted version of the same pipeline reads the same pages.
It runs in Docker on Render.
What happened
Settlements go from an atVenu PDF to a NetSuite import file.
Manhead is the same company as the side-by-side NetSuite migration, and I run 11 live systems for them today.
