Gaze tracking - calibrated
Compose
Eye dwell
Rest eyes on a target
84%
I would like a cup of tea
AI suggestions
SC
Sarah
caregiver
Listening - transcribing live
Input method
Dwell time
800 ms
Switch scan
1100 ms
Camera pointer
Off
No camera
Voice profile
My voice (cloned)
Trained from 2019 home video archive - 14 hrs source audio
Language model
Not loaded
Run model training to generate suggestions.
npm run train:model
Custom phrases
0 saved
Personal memory
Learning locally
Accepted suggestions are ranked higher on this device.
Reading speed