Case Study: Building Zalud, a Patient-Owned Health Records Platform

Zalud set out to do something that sounds simple and isn’t: give people real ownership of their medical history. The promise, “Tu historia clínica, con vos” (your medical history, with you), means sensitive health data has to be portable, private, and correct. That leaves no room for the shortcuts that are fine in a throwaway MVP and quietly turn into liabilities in a health product.

Fluxa Labs built the platform from scratch — architecture, backend, the admin panel, and the production setup — end to end, by a single senior engineer, from first commit to production in under eight weeks.

What we built

The product has two sides:

• A patient-facing app where people access and own their clinical history.
• An admin panel, built with React + Vite, for managing patients, clinical records, and the documents attached to them, with role-based access for the team running the platform.

The architecture, and why

For a small team building a health product, the right call was a lean, security-first stack with as little undifferentiated infrastructure to babysit as possible.

• Supabase as the backbone: managed Postgres, authentication, file storage, and, most importantly, row-level security (RLS). RLS lets you enforce data isolation inside the database itself. A patient can only ever read their own records, no matter what the application layer does. For health data, pushing access control down to the data layer is the difference between “we hope the API is right” and “the database guarantees it.”
• Vercel for hosting and deployment: preview deploys on every change, a global edge network, and no servers to maintain. The focus stays on the product instead of ops.
• React + Vite for the admin panel: a fast build and a tight feedback loop for the internal tooling that day-to-day operations lean on.

The discipline that matters here isn’t the logos in the stack. It’s the decisions behind them: modeling the data so privacy is structural, not bolted on, keeping the moving parts few, and leaning on managed services so a small team ships features instead of maintaining plumbing.

The result

Delivered in under eight weeks, by one engineer:

• 18 versioned database migrations — every schema change reviewable and reproducible
• 44 row-level-security policies — patient data isolation enforced inside the database, not just the API
• 9 serverless edge functions for the backend logic
• A full React + Vite admin panel with role-based access

For comparison: a typical agency quote for a build like this is a 3–4 person team over one to two quarters. Zalud got to market in a fraction of that time and cost — without trading away the security bar a health product demands.

Zalud launched as a privacy-first platform where patients genuinely own their records, on a foundation a small team can run and extend with confidence, not one that needs a rescue sprint twelve months later.

Hear it from Zalud’s CEO, Natalia Blanco:

The takeaway

You don’t need a big team to build a serious product. You need the right architecture, decided early, by someone who has carried systems into production before, so the decisions that are expensive to undo get made correctly the first time.

That’s the whole idea behind Fluxa Labs. If you’re building something where the technical bar matters, let’s talk.