n8n, Ollama & Docker: Extract PDF Data with Free Local AI

About this course

Stop copying data out of PDFs by hand. In this course you'll build a complete automation that reads PDF receipts and tickets, extracts the important fields with AI, and saves them into a clean, searchable data table, all running 100% locally on your own computer.

You'll use n8n to build the workflow, Ollama to run a free, open-source AI model on your machine, and Docker to run everything in a clean, repeatable setup. There are no API keys, no monthly subscriptions and no per-document charges. Your files never leave your computer, which makes this approach ideal for invoices, receipts, statements and other sensitive business documents.

The course is fully hands-on. You won't just watch slides; you'll build one real project from start to finish, with every lesson adding a new piece to the same working system.

You'll start by setting up Docker, n8n and Ollama, and choosing a local AI model that runs well on an ordinary laptop. Next, you'll extract structured data such as dates, names, locations and amounts from a single PDF, and learn how to make the AI return clean, reliable JSON instead of unpredictable text. You'll then design a data table and store every extracted record so it can be searched, sorted and totalled.

Once one file works, you'll scale up to processing a whole folder of PDFs automatically. Real documents are messy, so you'll learn to handle different layouts, strange file names and files that don't follow the expected pattern. You'll add validation to catch missing fields, bad dates and duplicate records, and you'll log failed files instead of losing them silently. You'll also build a cleaning step that standardises dates, currency and text into analysis-ready data.

Finally, you'll deploy the workflow so it runs on its own whenever a new file lands in your folder.

Read the full description on Udemy on Udemy (opens in a new tab)

Prices, discounts and availability are set by Udemy. We may earn a commission when you purchase through links on this site.