DEV Community

Mohamed Amer
Mohamed Amer

Posted on

I’m non-technical, tried to add RAG to my AI agent… and ended up building a tool that does that

I’m non-technical, tried to add RAG to my AI agent… and ended up building Lumine now I’m stuck — need your advice
A few months ago, I tried to build a simple AI customer support agent using n8n.

Sounds easy, right?
Then I realized: to make the answers actually useful, I needed RAG (Retrieval-Augmented Generation).
And that turned out to be much harder than I expected.

🧩 Why it was harder than it looked
I’m not a technical founder.
So I looked for existing RAG solutions, hoping for something “plug & play.”

Instead, I found:

Complex docs

Assumptions that you already have infra knowledge

Vendor lock-in everywhere

For someone who just wants to build fast, it felt… impossible.

⚙️ What I did next
It took time (and a lot of trial & error) to find the right technical partner.
Together, we built a tiny service to scratch our own itch:

Upload files

Get an endpoint

Done

We called it Lumine.

demo

🚀 Where Lumine is now
Right now Lumine is in beta:

You can upload your docs

You get an endpoint to query them

No forced vendor lock-in

Faster and simpler than what we tried before

Target users?
Honestly… still figuring it out. Could help:

SaaS builders

Automators

AI agent creators

🤔 Where I’m stuck
We have a small waitlist (about 20 people).
But no real users yet.

I’m not sure how to get:

The first few real users

Honest feedback

Ideas on how to position & market this

🧡 Why I’m writing this
If you’ve built something similar, or faced the same RAG pain, I’d love to hear:

How did you validate early?

Where did you find your first users?

What would you do if you were me?

Also: if anyone wants to try Lumine and share feedback, tell me and I’ll DM you access.

Not a pitch — a real question:
What would you do in my place?

rag #ai #startup #founderstory #api

Top comments (0)