Process Multi-Page Thai PDFs: Typhoon OCR + AI to Google Sheets

This workflow processes multi-page Thai PDF documents using Typhoon OCR, one of the most accurate tools for Thai text recognition. It starts by loading PDFs from a local folder, splits them into individual pages with pdfseparate and pdfinfo, runs OCR on each page via custom command execution, aggregates the extracted text, and employs an AI model (like GPT-4 or OpenRouter) to parse and structure key fields into JSON format. Finally, it converts the data into a tabular structure and appends one row per file to a Google Sheets spreadsheet.

Ideal for HR, payroll, and operations in Thai-speaking environments, it handles government forms, invoices, memos, and official letters effortlessly. Benefits include eliminating manual data entry, ensuring high OCR accuracy for Thai script, and enabling scalable automation without custom scripting.

Requires self-hosted n8n with Typhoon OCR Python package, poppler-utils, and local dependencies. Perfect for enterprises processing high volumes of Thai documents, saving hours of tedious work and reducing errors in payroll processing.