Reading an explanation is one thing. Hearing it, asking back, and hearing the answer is how most of us actually learn. Until now, a tutor that talks meant sending your voice to someone's server and paying by the minute. We built the Voice room the other way round: the talking happens on your device, which is why it is private, and why there is no meter on it.
Open any subject, tap Voice room, and ask a question the way you would ask a teacher across a desk. The tutor answers out loud, from your own material, the same notes and chapters your text chat already uses. You can interrupt, ask again, switch language mid-conversation, and keep going. Everything you say and everything it says is saved into that subject's chat, so nothing you learned by talking is lost.
What "on your device" means
The first time you open the room, it asks which voice you want and tells you exactly what it will download: a lighter voice at about 60 MB, or the best voice at 200 MB or 400 MB depending on the quality you pick. That download happens once and stays in your browser. From then on, every word the tutor speaks is generated by your own computer, on its graphics chip, in a fraction of the time it takes to say it.
Listening works the same way: the room uses the speech recognition already built into your browser to turn what you say into text. No recording of your voice is ever uploaded to iTutor, and none is stored anywhere.
What does reach iTutor is the text of your question, the same as when you type in chat. That is what the tutor needs to find the right page of your material and write an answer. It is also all it gets.
Why that makes it unlimited
Every spoken sentence a voice tutor produces on a server costs its maker something, so voice tutors count minutes. When the voice is made on your device, there is nothing for us to meter. So there is no monthly allowance, no "you have three minutes left", no timer in the corner. Talk through a whole chapter. Come back tomorrow and do it again.
We think this is the honest way to give students a talking tutor. A voice that stops after ten minutes teaches you to ration your questions, which is the opposite of what a tutor is for.
Who this is for
This is the whole point of the room. A tutor you can talk to, that costs nothing per minute and keeps your voice to yourself, changes who gets to learn this way.
Ten languages, one voice
The best voice speaks English, Arabic, French, Spanish, German, Portuguese, Italian, Dutch, Turkish and Indonesian. Arabic is a first-class citizen here, not an afterthought: it is the same voice, the same quality, the same patience. For Malay and Urdu the room uses your device's own voice, with nothing to download.
Phones and laptops
On a laptop or desktop with a graphics chip, you get the best voice. On a phone, the room offers the lighter voice instead, which is quick to download and easy on the battery, and the device's own voice as an alternative. If your device turns out to be slow for the voice you picked, the room says so and offers a lighter one; it never leaves you waiting in silence.
Your choices stay on your device
The voice you chose, the quality tier and the download all live in your browser, on your device. Nothing about them is sent to us. If you ever want to start clean, there is a Reset voice choices on this device button in the room's settings, and clearing your browser's site data removes the downloaded voice entirely.
Where to find it
Open a subject and look for Voice room under the subject's tools. Pick a voice, let it download once, tap Start talking, and ask your first question out loud. If you would rather read than listen at any point, the same conversation is waiting for you in the text chat.
We built iTutor so that a student anywhere could have a patient tutor for free. A tutor you can talk to, whose voice lives on your own device and answers for as long as you need, is what we always meant by that.