Skip to content

Projects

FONEMA

English phonetics, with voice processed locally

The FONEMA sign-in page: the name in wide black letters, the transcription /fo.ˈne.ma/ in blue, a sentence explaining the app and a field to sign in with just an email.
Year
2026, ongoing
Role
Sole author. My own product.
With
Next.js, React, TypeScript, Supabase, Docker, Vertex AI
Links
Open FONEMA

What it is

FONEMA is an English phonetics lab for Spanish speakers: minimal pairs (ship and sheep), dictation in the phonetic alphabet, IPA reading and review cards. You sign in with just your email, no password.

What makes it different

Voice goes local first: speech synthesis and recognition run on my own hardware and don’t depend on someone else’s service. And review isn’t a generic list: it comes from what the person heard that day.

How it’s built

Next.js and React on Supabase, packaged in Docker. Vertex AI models come in only where local voice falls short.

All projects