Case study · Project 02 / 04
Invoice Extraction Pipeline
Drop an invoice into Drive, get structured data back out. OCR and extraction on Groq, no database in the loop.
- Role
- Automation design, build
- Stack
- n8n, Groq API, OCR, Google Drive
What is Invoice Extraction Pipeline?
The Invoice Extraction Pipeline is an n8n workflow that turns invoices into structured data. An invoice dropped into a Google Drive folder is read with OCR, extracted by an LLM on the Groq API, validated, and written back to Drive as structured data. There is no database in the loop.
What problem does it solve?
Invoices arrive as documents, and the data inside gets re-typed by hand. The pipeline removes that step: the document goes in, and the invoice number, vendor, line items, tax and total come out as consistent, structured fields.
How does it work?
Four stages run in order. Drive picks up the invoice, Groq reads and extracts it, a validation step checks and shapes the fields, and the result is written back to Drive. n8n orchestrates every hand-off, so the whole flow needs no database.
- 01Input: DriveAn invoice lands in a Drive folder and n8n picks it up.
- 02Process: Groq OCR and extractionThe document is read and an LLM extracts the invoice number, vendor, line items, tax and total.
- 03Transform: ValidateFields are checked and shaped into one consistent structure.
- 04Output: DriveStructured data is written back to Drive. No database in the loop.
What is it built with?
It is built with n8n for orchestration, the Groq API for OCR and LLM extraction, and Google Drive for both input and output.
What does it show?
Automation depth: a production-shaped document pipeline built from n8n and an LLM API, running without a database.