Build a Free RAG system with N8N

RAG Time ! Yesterday, Udemy had their Discount Day again. I am hooked to that. As usual, the platform offered countless courses at discounted prices, and this time was no different. As I was browsing through the offers, one specific course caught my eye: a course on N8N, a powerful and flexible workflow automation platform. The Udemy N8N course description immediately grabbed my attention, especially since it mentioned something about RAG (Retrieval-Augmented Generation). Given that I had been exploring RAG systems lately, I knew I had to get this course.

El Cheapo !?

The price ? Just 14 dollars for the discounted course, a bargain in my opinion.

I am quite the El Cheapo, I love services with free tiers. I try 5 services a week, on average, and I don’t feel like paying for them all, where I don’t use them in general after trying them for a week. I just want to see what a service is capable of, what the quality of their production is, what their use is in the general workflow. I don’t have the money to pay for every service, I don’t even have a credit card. So I am basically El Cheapo, you know, “Free” Willy ! Give me a free tier ! Same with Udemy courses, I buy one every month on average, I can wait for the Discount Days. El Cheapo loves Discount Days !

I added it to my cart and purchased it. It turned out to be an excellent decision, as it gave me exactly what I was looking for. The course walked me through the process of setting up a simple RAG system, and even better yet : the course provided a ready-made template that I could use straight away. This gave me a solid foundation to build upon.

However, there was one small issue. The solution presented in the course relied on OpenAI, a paid service. I don’t have a credit card. So I tried some other services looking for services with a free tier. My goal was to set up a RAG system that could operate for free, at least within the free tiers of certain services. So, I decided to tweak the template and make some adjustments.

After some miserable fails, I managed to get the system working with three tools with a free tier: Qdrant, HuggingFace, and Groq. Of course, this modification meant that the system now had some limitations compared to using a paid service. But the good news was that I had a functioning RAG solution that operated entirely for free (within the free limits of these services).

The result? A local, free AI-powered RAG system. While it may not have the power or scalability of the paid version, it’s a nice solution for small-scale use or for experimenting without incurring any costs.

I tested it on the board meeting notes of our charity and it works like a charm. It is excellent for that kind of thing. You can chat with your own archive.

N8N RAG system template

You can download this N8N Free Rag system template.

You can unzip and load it in N8N. Once you setup the credentials, you can easily load your own Word documents into an AI-powered Q&A system. The process is simple and efficient.

In the first step, you can upload a Word document directly from your Google Drive into your RAG (Retrieval-Augmented Generation) database. This allows you to store and organize your own content for retrieval-based AI interactions.

Then, in step two, the magic happens. Using Groq, you can engage in meaningful conversations based on the content of your uploaded documents. This means you’re not just chatting with an AI in a generic way; you’re actually able to query the system using the information from your own files.

Notes

Disclaimer: “It works on my machine” 😊. I’ve done my best to ensure that this N8N Free RAG template runs smoothly, it works like a charm on my laptop. And I trust it will work just as well on yours. That said, one of the more time-consuming steps is setting up the necessary credentials. Integrating services like Google Drive requires obtaining API keys and handling Google OAuth consent to access your documents. While the process can seem a bit tedious, it’s essential to ensure secure and seamless access to your data. Once you’ve got those credentials in place, the rest of the system falls into place effortlessly.

I hope you find this setup useful for your own AI-powered document workflows. Whether you’re using it to retrieve insights from your own files or building a custom knowledge assistant, this free RAG template provides a solid starting point. Good luck, and I hope you enjoy building and experimenting with this system as much as I did!