Voice
Gray has two voice layers. Knowing which is which explains the plans.
On-device voice — unlimited, free
Section titled “On-device voice — unlimited, free”Your speech is transcribed on the phone (the platform’s own on-device speech recognition) and the text goes straight to your box. No audio leaves your hands, there is nothing to meter, and it works the same on every plan. This is Gray’s default voice.
Replies are written into the feed as Gray works. Spoken replies are part of premium voice — the standard path is voice-in, text-out.
Premium voice — metered, brokered
Section titled “Premium voice — metered, brokered”Premium voice is the fluid, low-latency conversational layer (it is also the lifeline that can talk you through connect/disconnect problems). It uses a realtime speech model brokered through your box: the box mints a short-lived session token and your phone streams audio directly to the voice provider for the length of that session.
Premium voice carries a monthly allowance that depends on your plan. When the month’s allowance runs out, Gray says so and falls back to on-device voice — you are never cut off from using Gray.
Voice speed
Section titled “Voice speed”Me → Voice speed is a continuous slider; drag to preview, release to save.
Languages
Section titled “Languages”Me → Language offers the box’s supported set (English, Spanish, Indonesian, Vietnamese, Filipino, Thai, Arabic) and steers premium-voice transcription, the voice persona, and the language Gray replies in. On-device transcription currently listens in English (US).