An automation that reads invoices and receipts — PDF or photo — and turns them into clean spreadsheet rows, with zero manual typing.

The Problem

Small businesses lose hours every week manually typing invoice and receipt data into spreadsheets. It's repetitive, error-prone, and gets worse when receipts come in mixed formats — some as scanned PDFs, others as phone photos with no consistent layout. Duplicate entries are common when the same file gets processed more than once.

The Solution

I built a workflow that watches a Google Drive folder for new invoice or receipt files. It automatically detects whether the file is a PDF or an image, routes it through the right extraction path, and uses AI to pull out the vendor, date, total amount, currency, and line items — then writes a clean row into a Google Sheet.

If the same file is ever processed again, the workflow updates the existing row instead of creating a duplicate, using the file's unique Google Drive ID to match records.

How it works

1. Watch folder — a Google Drive trigger detects any new file dropped into an invoices folder
2. Check file type — the workflow branches based on whether the file is a PDF or an image
3. Extract content
   • PDFs: raw text is extracted directly from the document
   • Images: the file is encoded and sent to a vision-capable AI model
4. AI extraction — in both cases, the AI reads the content and returns structured data: vendor, date, total amount, currency, and itemized line items
5. Save to spreadsheet — the data lands in Google Sheets, matched against the source file's unique ID so reprocessing never creates duplicates

What this saves

• Eliminates manual data entry per invoice
• Removes typos and missed line items from manual transcription
• Handles two different file formats through one consistent process
• Prevents duplicate bookkeeping entries automatically

Tools used

n8n · OpenAI (GPT-4o mini, vision + text) · Google Drive · Google Sheets