About Ookie
Everyone should get to hear how good they can sound
Ookie plays any sentence in your own voice, the way a native speaker would say it. Then you copy yourself.
Built on three ideas
Nothing here is a trick
Three ideas that already work in language teaching, put together in one short loop.
Shadowing
Copying a voice you can hear is how most of us learn an accent. Research links it to faster, more confident speech.
The golden speaker
A model voice has to be one you want to sound like. Research links that to noticing your own mistakes sooner.
Your ideal self
Hearing the version of you that already speaks the language makes the next sentence easier to say.
The research behind these ideas
Shadowing and pronunciation feedback: Whitworth 2024, Probst 2002, Ding 2019. Ideal-self and possible-self models: Al-Hoorie 2018, Dowrick 1999.
Under the hood
What actually happens
Three things run behind every sentence. None of them need a lesson plan.
Voice cloning, once
Two short sentences are enough. One recording becomes the voice you hear in every language you practice.
A native accent, every sentence
Whatever you highlight, you hear it the way a native speaker would say it — not in a generic robot voice.
Speech scoring, every take
Pronunciation, fluency and intonation are scored on the take you just recorded, and turned into one thing to fix.
Your voice is yours
Private by default
Your voice is private to you · Delete it anytime. Your recordings stay on your computer; only scores and your streak are saved to your account.
Hear how good you already sound
Start with the next sentence you read, not with a lesson plan.