Open-Source Invoice Extraction API with OCR and ValidationOpen-Source Invoice Extraction API with OCR and Validation
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
I built an invoice extraction API that costs nothing to run. You upload a PDF or image, and it returns clean structured data: vendor, invoice number, dates, line items, subtotal, tax, total — every field with its own confidence score.
Under the hood it runs PaddleOCR on a regular CPU (Tesseract as backup), then a rule-based parser pulls out the fields. Before anything gets accepted, an arithmetic check makes sure subtotal plus tax actually equals the total — that catches the small digit errors OCR often makes. If something looks uncertain, it goes to a human review queue instead of failing quietly, and every correction is saved in an append-only log so you can always see what changed.
No paid vision APIs, no GPU needed. The whole stack is open-source, it ships in Docker, and it's covered by 28 unit and 4 smoke tests.
Post image
Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started