# Meditation Studio audio

The four meditation scripts and eight atmosphere/harmony layers are original
AutoGeniusOS material authored with AI assistance for Jody Sachse. The weather
textures are synthesized compositions, not field recordings or licensed music.
They contain no binaural beat or rhythmic entrainment target.

The optional spoken guidance is computer-generated narration rendered with
[Piper](https://github.com/OHF-Voice/piper1-gpl) 1.8.0 and the
[LJ Speech voice](https://huggingface.co/rhasspy/piper-voices/tree/main/en/en_US/ljspeech/high).
The [LJ Speech dataset](https://keithito.com/LJ-Speech-Dataset/) is public domain.
The voice's model card is retained beside this file. Its exact source revision
and downloaded file hashes are recorded in the audio manifest.

Piper is an authoring tool; its engine and voice model are not distributed in
the app. Only the finished narration recordings are bundled. No speech service,
microphone, voice upload, or AI account is needed during playback. This voice
is a narration instrument and does not represent Kai or a live AI companion.

The immutable asset manifest contains codec-independent source levels, durations,
loop-boundary checks, cue timing, and final file hashes. Authoring scripts live
in `mobile/scripts`; the APK build verifies the delivered files instead of
regenerating voices or soundscapes each time.