Where simple ideas become potential products
I'm the founder of Zaguán, an LLM proxy with a powerful dashboard that helps teams manage their AI interactions. I've been in IT since the mid-90s, started using Linux in 1996, and have worn many hats throughout my career.
The idea for Zaguan Labs came from a simple realization: I'd just completed Zaguán Orchestrator after building a successful Python proof-of-concept. That moment made me realize something important—a simple PoC could potentially become a viable product.
So I asked myself: "What if I shared all my experiments publicly?"
Zaguan Labs isn't just about prompt engineering or AI security—it's about trying new things. Finding new ways to use AI. Pushing boundaries. Each experiment here is a potential solution to a problem I'm exploring or a wild idea I want to test.
I'm always asking: "How far can I take this?" Every project you see here uses Zaguán as the backend for accessing AI models. It's my testing ground, and now it's yours too.
An LLM proxy with a comprehensive dashboard for managing AI interactions at scale.
Coming soon—born from a successful PoC, now being developed into a full product.
This playground. Where the next product might be hiding in plain sight.
Here's the thing: I'm a one-man team right now. That means every experiment you see here is something I built because I thought it was interesting or useful. But I can't develop everything into a full product.
That's where you come in.
If an experiment gets attention—stars on GitHub, discussions, people actually using it—that tells me it's worth developing further. A simple PoC that resonates with people could become the next member of the ecosystem.
If you find an experiment useful or interesting, give it a star. It's the simplest way to show support.
Have ideas for improvements? Found a bug? Want to suggest a feature? Open an issue or discussion.
If something here solves a problem for you, share it with others who might benefit.
See room for improvement? Pull requests are always welcome. Let's build something better together.
I don't have a fixed schedule for publishing new experiments. They'll appear when inspiration strikes or when I solve an interesting problem. Some will be small utilities, others might be ambitious prototypes.
What they all have in common: they're real, working code that you can use today. No vaporware, no promises—just experiments you can run, modify, and learn from.
Who knows? The next experiment might just become the product you've been waiting for.