FluencyAI device probe

Six checks that decide whether a browser version of FluencyAI can carry voice. Run every one, then send me the results.

Running as checking…

Run this twice: once in a normal Safari tab, then again after Share → Add to Home Screen. Speech recognition and notifications behave differently in each, and that difference decides the architecture.

Environment

What this device and browser actually are.

reading

…

Ana's voice

Whether Ana can speak her replies out loud.

not run

Tap to hear a test sentence and list the installed English voices.

Voice input — English

The core interaction: speaking to Ana instead of typing.

not run

Tap, allow the microphone, then say: "I am agree with you"

Voice input — Hindi

Needed for the Translate to Practice module.

not run

Say any short Hindi sentence. An error code here is still a useful answer.

Voice input — Marathi

The second Translate to Practice language.

not run

Say any short Marathi sentence.

Microphone capture

Raw audio for pronunciation scoring. The browser records, Azure scores.

not run

Records a short clip and plays it back. If playback works, pronunciation scoring works.

Reminders

Daily practice nudges and flashcards-due alerts.

not run

Checks the notification and push APIs, then asks permission.