Real voices, not TTS
Clips come from Mexican speakers in Mozilla Common Voice, filtered for validated, dictation-length recordings.
¿cómo dice?
Train your ear for
Mexican Spanish.
Listen to a real human speaker, type what you hear, see exactly where you slipped.
Built on Mozilla Common Voice. Memory via FSRS. No TTS, no fake cheerfulness.
What's in the box
Type on desktop
Character-level diff shows which accent you dropped and which consonant you missed, so you stop repeating the same mistake.
Spaced repetition
FSRS reschedules phrases based on your grade so you work the ones that actually need work.
Speak on mobile
Mic mode uses the browser's own recognizer in es-MX. For when your thumbs are not on a keyboard.