Vernato

Vernato comes from my own frustration with pronunciation practice. I wanted feedback that gave me something specific to try again. I’m building it as an iOS app where a conversation can lead into a short pronunciation exercise, followed by a closer look at the words and sounds that need attention.

A score needs somewhere to go

A score can tell you whether an attempt was good or bad, but by itself, it is not very useful. I wanted Vernato to go further than that, so each assessment breaks a sentence down into the words, syllables, and sounds that need attention, while still letting you come back to past attempts later. A lot of the work has gone into making that detail feel useful rather than overwhelming, because I believe feedback only matters if you know what to try differently next time.

The parts behind the practice

Building Vernato has been a long but worthwhile process. It started as a web app, but I eventually realized pronunciation practice made much more sense on a phone, especially if it could work whenever you had a few minutes and continue offline. That pushed me toward rebuilding it for iOS, using my Figma prototypes alongside Codex and Claude Code as I learned a very different development workflow.

The harder part has been working with Azure’s speech and voice APIs, handling authentication and privacy, caching results, enforcing usage limits, and deciding what actually needs to reach a server as I am trying to keep the infrastructure within free-tier limits.

It is still in development. I’m proud of how much of it I’ve changed since the first version, and I’m still working on it.