Águila

The problem

A few hundred students leave the same building at the same time every day and eat within ten blocks of it. The restaurants in those ten blocks want them. Neither side has any way of finding the other, so the whole thing runs on word of mouth and whoever happens to have a sign out.

Where the idea came from

Not from me. Gabriel Slotnisky had been running a version of it alone since the start of the year, called MorfaloYa, and it was not moving. We had met at a talk, and he wrote to ask whether I wanted to do it properly.

So we split it the obvious way: I wrote the code, he sold, and Nicolás Krymkiewicz came in for the design and the promotion banners. It stopped being an idea and started being a thing within a few weeks, which is mostly a fact about Gabriel selling.

What it was

A Next.js app with two halves that shared a database:

• The public side — a directory of venues near school, each with its current promotions, at /r/[restaurant]. Internationalised from the start through a [lang] segment, because Águila was meant to work past one school.
• The dashboard — where a restaurant managed itself: its own page, its categories, its discounts. Auth with an email sign-in and callback flow, so an owner could log in and change a promotion without going through us.

That last part was the whole product thesis. A directory somebody else has to update is a spreadsheet with extra steps; the thing worth building was the venue changing its own offer at four in the afternoon because lunch went badly.

What happened

We reached ten restaurants paying us, several with promotions exclusive to the platform. It worked. It was small, and it worked.

Then Gabriel said he would rather put himself one hundred percent into studying for his university entrance exam. He offered to hand it to Nicolás and me to carry on. The three of us talked it through and agreed together to leave it there.

We refunded every restaurant. None of them asked. Several had months left on what they had paid for and would have let it go.

Then we opened the source. If we were not going to run it, the work may as well be readable.

Why that is on my portfolio

Because the alternative was keeping money for a service we had stopped providing, from ten small businesses, on the grounds that the contract did not say we had to give it back.

It cost us the only revenue any of us had ever made. I would do it again without thinking about it, and I would rather be the kind of person who finds that obvious than the kind who has to decide.

What I would do differently

Agree what happens when someone leaves, before selling the first month.

We never had that conversation, so when it arrived it had no framework, and the path of least resistance was to stop. It is possible Águila should have closed anyway — I am not sure it was a business rather than a good favour to ten restaurants. But we did not decide that; we defaulted into it.

That gap is exactly what the founders agreement at Restoman exists to close, and I would not have known to write one if this had gone well.

Note

aguila.lat expired in September 2026 and was not renewed, so there is nothing left to link to. The code is open source and still up.

The one picture of Águila that survives is the mockup on the work index, and it survives because it was made to be shared rather than to be kept. I did not take screenshots while it was running, so the dashboard, the rest of the venue pages and most of the banners Nicolás made went with the domain. If you build something that might not last, photograph it while it works.