This n8n workflow automates the processing of bank statements, transforming PDFs into structured data using AI-driven vision and language models. Starting with a manual trigger, the workflow downloads a bank statement PDF from Google Drive, then converts each page into images via an external API. These images are unzipped into individual files, resized to optimize AI processing, and transcribed into markdown format using a vision-enhanced language model. The transcribed text is then analyzed with a Google Gemini chat model to extract specific deposit table rows, facilitating financial data analysis automatically. This workflow is ideal for accounting, auditing, and financial analysis, especially for scanned PDFs or complex document layouts that traditional OCR struggles with, providing a seamless, automated data extraction pipeline.

Node Count	11 – 20 Nodes
Nodes Used	@n8n/n8n-nodes-langchain.chainLlm, @n8n/n8n-nodes-langchain.informationExtractor, @n8n/n8n-nodes-langchain.lmChatGoogleGemini, aggregate, code, compression, editImage, googleDrive, httpRequest, manualTrigger, sort, stickyNote