You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Mobile voice settings: show which on-device speech engine runs
#33710
In mobile Settings → Voice, the This device option says only "Audio stays on this device and is transcribed here." It does not say which engine transcribes. The environment rows below it already show model state (for example "No transcription model installed"), so on-device is the only destination with no model information.
This causes real confusion on Android. T3 Code uses Google's ML Kit GenAI Speech Recognition and selects Advanced mode (Gemini Nano, newer phones only) or falls back to Basic. Most phones, such as a Pixel 7 Pro, get Basic, which is noticeably less accurate. Users cannot see this, and some assume the option uses their keyboard's voice typing or their default voice input app. The app says "Your keyboard's voice typing is separate" only when on-device dictation is unavailable.
Proposal
Add one line under This device, plus a short section footer:
Android: Google on-device speech · Basic · English (United States), or · Downloads on first use when the model is not on the phone yet. The footer explains Basic and Advanced, and says the keyboard's voice typing is separate.
iPhone: Apple on-device speech. The language is chosen only when dictation starts, so it is not shown.
The Android availability probe that already runs on app start and foreground returns the mode it would pick and whether its model is downloaded. It uses the same order as prepare and never downloads anything.
Before
After
Mockup of all states (Basic, Advanced, not downloaded, iPhone):
An implementation PR is ready and linked below, so it can be merged directly if this direction is approved.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Problem
In mobile Settings → Voice, the This device option says only "Audio stays on this device and is transcribed here." It does not say which engine transcribes. The environment rows below it already show model state (for example "No transcription model installed"), so on-device is the only destination with no model information.
This causes real confusion on Android. T3 Code uses Google's ML Kit GenAI Speech Recognition and selects Advanced mode (Gemini Nano, newer phones only) or falls back to Basic. Most phones, such as a Pixel 7 Pro, get Basic, which is noticeably less accurate. Users cannot see this, and some assume the option uses their keyboard's voice typing or their default voice input app. The app says "Your keyboard's voice typing is separate" only when on-device dictation is unavailable.
Proposal
Add one line under This device, plus a short section footer:
Google on-device speech · Basic · English (United States), or· Downloads on first usewhen the model is not on the phone yet. The footer explains Basic and Advanced, and says the keyboard's voice typing is separate.Apple on-device speech. The language is chosen only when dictation starts, so it is not shown.The Android availability probe that already runs on app start and foreground returns the mode it would pick and whether its model is downloaded. It uses the same order as
prepareand never downloads anything.Mockup of all states (Basic, Advanced, not downloaded, iPhone):
An implementation PR is ready and linked below, so it can be merged directly if this direction is approved.
All reactions