app-audio-sandbox
v0.1.2
Published
A Playwright-powered silent audio sandbox for app voice flows.
Readme
App Audio Sandbox
app-audio-sandbox is a Playwright-powered CLI for testing app voice and
audio flows without playing sound through your speakers.
It launches or attaches to an app, injects browser audio/STT/TTS instrumentation, records events, captures TTS audio artifacts, and loops captured app audio back into the sandbox microphone while keeping playback muted by default.
Install
npm install -g app-audio-sandboxUsage
app-audio-sandbox list
app-audio-sandbox run --scenario nudge-voice-cutoff
app-audio-sandbox run --scenario nudge-voice-cutoff --record-audio
app-audio-sandbox run --scenario nudge-voice-cutoff --record-audio --real-tts
app-audio-sandbox run --scenario nudge-voice-cutoff --audible--record-audio captures audio source files without speaker playback. --audible
is the only mode that allows sound to reach the speaker. --real-tts lets the
target app's TTS endpoint generate listenable audio instead of the deterministic
mock audio used by default. Recorded Nudge runs also include review.html, which
plays captured app audio and browser-friendly scripted STT WAV inputs on the same
timeline so cutoff causes are audible during review.
Built-In Scenarios
nudge-voice-cutoffnudge-echo-filternudge-native-tts-routingnudge-stt-warmup
Each run writes an artifact folder with events.json, a Playwright trace, video,
final screenshot, review.html when a review timeline is available, and captured
audio files when recording is enabled.
