Can n8n handle RAG and heavy PDF parsing natively?
Hi everyone. I'm building an AI Contract & ToR Triage Agent for my team to automatically extract SLAs, risks, and technical requirements from public bidding documents (30+ pages, messy tables).
I originally planned an architecture using an Azure VM with Python and Docker for the heavy lifting (OCR, text cleaning, chunking). However, since I sit in a business team outside of the IT department, my request for the VM was denied due to strict security policies and budget constraints. Every new piece of infrastructure is seen as an unnecessary cost.
Can this be done entirely within n8n? The pipeline I need: Ingest PDF -> OCR -> Clean & Chunk text -> Embed -> Query LLM
- How do you handle OCR and chunking of 30+ page PDFs purely within n8n? Are there reliable community nodes or cheap/serverless APIs you recommend calling from n8n to offload this?
- Without a dedicated Docker host for something like Qdrant, what is the most cost-effective Vector DB approach that integrates smoothly with n8n's AI nodes?
Any advice on pushing n8n to its absolute limits for Document Intelligence would be amazing!