Email is the fastest way to reach me directly — I read every message and usually reply within a few days.
Check that Nabi has microphone and speech-recognition permission — on iOS, go to Settings → Nabi and confirm both are allowed. Recognition runs through your phone's own speech engine, so the first use may need a one-time language-pack download over an internet connection. Speak only after the mic button shows it's listening.
No. Speech recognition runs entirely on your device. Each shadowing attempt is cached locally so you can play it back and compare it with the native recording, then it's overwritten by your next attempt. Audio is never sent to our servers — see the Privacy Policy for the full picture.
Progress is tied to the specific Apple or Google account you signed in with. Double-check you're signed in with the same provider and account you originally used — lessons, streak, and review schedule don't transfer between different sign-in methods or accounts.
In the app, go to Settings → Privacy & data → Delete account. This removes your profile, progress, streak, and review schedule immediately and can't be undone. Use "Download my data" first if you want a copy before deleting.
The tutor is a language model, and language models are sometimes confidently wrong — especially on nuance, formality, or regional usage. Treat its corrections as practice, not authority, and don't rely on it for anything where being wrong has real consequences.
Yes — no ads, no subscription, no lesson behind a paywall. The free tutor runs on a shared AI tier; if you'd rather your conversations not go through that shared tier, add your own provider key in Settings → Use my own AI key and requests go straight from your device under your own account.
Email support with your device model, OS version, and what you were doing right before it happened — a screenshot or screen recording helps a lot.