Extract bank statement transactions from PDFs to Google Sheets with Google Drive and Gemini

Quick overview

Accountants, freelancers and small businesses in Latin America who copy bank statement transactions by hand.

How it works

1. Drop a statement PDF into a Google Drive folder.
2. The workflow downloads it and extracts the text.
3. Google Gemini returns every transaction as JSON (date, description, amount, debit/credit, balance), handling both 1.234,56 and 1,234.56 formats.
4. Each transaction is appended as a row in Google Sheets.
5. Always check totals against the original statement.

Setup

1. Add Google Drive, Google Sheets and Google Gemini credentials (free API key from https://aistudio.google.com/apikey).
2. Pick the folder in the trigger.
3. Create a sheet with headers: Fecha | Descripcion | Importe | Tipo | Saldo | Archivo, and select it in the Sheets node.
4. Text-based PDFs only (run OCR first for scans).
5. Activate the workflow once the credentials are added.

Requirements

• n8n (cloud or self-hosted) and a Google account. The default model (gemini-3.5-flash-lite) is on the Gemini API free tier, so it runs at no cost within the free rate limits (free-tier data may be used by Google to improve its products; switch to a paid key for sensitive statements).

Customization

• Change the model or prompt in "IA: leer movimientos"
• Swap the trigger for Gmail attachments
• Add a category column to the prompt

Additional info

Text-based PDFs only (run OCR first for scans). Always check totals against the original statement.

Made by Agente Diario — daily AI & agents newsletter in Spanish: https://agentediario.beehiiv.com/subscribe?utm_source=n8n