ESP32 Robot: Phone Voice and Dual-Speaker Replies
Android 1.4.19 adds phone speech input and Xiaozhi replies on both the phone and robot.
The Android app now sends phone speech to Xiaozhi and plays replies on both the phone and robot. The owner confirmed this workflow on November 5, 2026.
This update pairs Android 1.4.19-phone-audio-dev (build 27) with firmware 2.4.47-phone-audio-dev. It remains a development build with the existing project signing certificate.
Open Xiaozhi in the controller, hold the talk button, speak to the phone, then release and wait. Keep the voice window open for replies. Phone playback follows media volume.
The voice connection uses a paired device certificate and a signed voice permission token. Bounded audio queues keep slow phone playback from blocking the robot speaker.
Slow, medium and fast gears live in Settings. The right stick locks rotation and camera pitch to separate axes.
TF-card remote codes, learned keys and direct infrared commands share the remote interface. Actual air-conditioner response still needs on-site testing.
Phone input and dual-speaker replies were confirmed. Long sessions, other networks, room echo and playback timing still need broader testing. The earlier reboot investigation is not declared resolved by this release.