Vitrifarm: why a price comparator had to become a pharmacy system first

Our first venture is in the world. Vitrifarm lets anyone in Venezuela search for a medicine and see what it costs at the pharmacies around them — by brand, by active ingredient, or by the symptom they actually typed. At the time of writing it carries 19,502 products across nine pharmacies in six cities, with stock status on every listing.

That is the part you can see. The part worth writing about is what had to be built underneath it, because the comparison engine was never the hard problem.

A price comparator is only as honest as its inventory#

Comparing prices is a sort operation. Any engineer can write it in an afternoon. The thing that decides whether a comparator is useful or worthless is the freshness of what it is sorting.

In retail, stale data is an annoyance — you drive over and the sneakers are gone. In pharmacy it is something else. Someone searching for an antibiotic at nine at night, in a city where that antibiotic is intermittently available, is not comparison shopping. They are trying to find out whether the thing exists within reach tonight. A listing that says Disponible when the shelf is empty does not just waste a trip. It costs someone a dose.

So the real question at the start was not how do we compare prices. It was how does accurate inventory from independent pharmacies end up in a database, every day, without anyone being paid to retype it.

Every honest answer to that question stops being a consumer product.

The supply side had to come first#

Scraping was never on the table — the data does not exist online to scrape. A feed integration assumes the pharmacies already run software that can emit one. The stock on those shelves lived, for the most part, in a notebook and a spreadsheet.

The way through was to build the thing the pharmacy actually wanted and let the inventory fall out of it as a by-product. So Vitrifarm's second product is a full administrative system for pharmacies: inventory that updates without typing product by product, a point of sale that charges the way payment actually works in Venezuela, online orders, every local payment method in one panel, delivery the pharmacy controls itself, sales reporting, customer reviews, automatic sync, and multiple branches under one account.

Read that list again and notice what it is not. None of it is in service of the comparator. It is a standalone business system that a pharmacy would pay for on its own terms, and the pitch to the owner is not help us populate a marketplace. It is the notebook and the Excel don't cut it anymore.

The marketplace is what you get for free once the pharmacy is running its day on your software. The inventory is live because it is the same inventory the cashier is selling from — not a nightly export, not a good-faith estimate. That is the only version of this product that can be trusted with a question about medicine.

This is what we mean in our own materials by a marketplace with a real supply story. Most two-sided products fail on the side that has nothing to gain. Vitrifarm's answer was to make that side the product.

Then a third side appeared#

Once pharmacies were running on it, a third participant turned out to be sitting right next to the problem: the prescriber.

A paper prescription is a dead end. The patient cannot read it, cannot verify it, and has no idea which pharmacy nearby can actually fill it. So Vitrifarm issues digital prescriptions — doctors register with their licence, the credential gets validated, they upload their signature and logo, and they issue and share from there. The patient receives something they can read and, because the catalogue is already underneath, a prescription that tells them where it can be filled. The prescriber gets anti-counterfeiting verification, clinical alerts while writing, and a patient agenda.

That is three audiences — patient, pharmacy, prescriber — and each one closes a loop the other two left open. A comparator alone would have been a nice utility. The three together are infrastructure.

Built for Venezuela, not translated for it#

Two details are easy to skim past and are most of the engineering.

Every price is shown in dollars and bolívares at the same time. Not a currency toggle in a settings menu — both figures, on every listing, always. That is not a localisation feature. It is how prices are quoted, negotiated and understood in the country, and a product that made you pick one would be describing a place that does not exist.

Every local payment method lands in one panel. Anyone who has built checkout in Venezuela knows this is not a Stripe integration with the locale changed. It is the actual texture of the market, handled rather than abstracted away.

Products built elsewhere and shipped into markets like this one usually get these exactly backwards: the currency is a dropdown and the payments are a single foreign rail. Building from inside the constraint produces different software.

Six cities, not a country#

Vitrifarm launched across Coloncito, La Fría, La Grita, Palogordo, San Cristóbal and San Juan de Colón — six cities, one state, Táchira.

The restraint there is deliberate. A comparator covering all of Venezuela with three pharmacies per city is a demo. A comparator covering one state densely enough that the nearest listing is genuinely the nearest pharmacy is a product someone uses twice. Density before reach: the second search is what tells you whether you have built anything.

What this one taught us#

Three things we are carrying into the next venture.

The unglamorous side of a marketplace is the whole company. The time went into the pharmacy system, which no consumer will ever see, and that is the asset. We would now be sceptical of any two-sided idea where the supply side is treated as a business-development problem rather than a product one.

Constraints are specification, not obstacle. Dual currency and fragmented payments were not things to work around. They produced a product that is genuinely hard to copy from outside the market.

Ship narrow and real. Six cities with accurate stock beats a national map with guesses, and it was in front of actual users months earlier.

Vitrifarm is live and open to anyone at vitrifarm.app. Every piece of it started where all of our work starts — as a hunch, with no product and no deck. If you are sitting at that stage now, tell us the shortest version of it.