The cloud sends text and the device speaks it with TTS (system engines: Windows SAPI / macOS/iOS AVSpeech / Android TTS; an optional bundled offline engine "sherpa" downloads a ~51MB model on first use). Multiple utterances play FIFO.
Samples
Fields
FieldNotes
textText to speak. FIFO queue; queue limit is 20 items
repeatTimes to repeat this item (default 1, max 99); the next item starts only after all repeats
langLanguage (zh/en); defaults to system language
volume0~100, default 60
rateRate factor 0.5~2.0, default 1.0
engineEngine: system (OS TTS) / sherpa (bundled offline engine, works without network)
stop / pause / resume / clearQueue control: stop the current item / pause / resume / stop & clear all

Scope: lang/volume/rate on a play command (with text) apply to that item only; a settings-only command (no text) updates and persists them (synced to the cloud config items speech_lang/speech_volume/speech_rate).

Replies & Progress Reports
  • Immediate reply: {"ret":"ok"} on success; queue full when full; engine not ready when no TTS engine is available; invalid param: ... for bad parameters
  • Progress reports (async, order/client, carrying the original mid): item started {"state":{"status":"1"},"progress":"1/2"}; item finished {"state":{"status":"0"},"progress":"2/2"}; items terminated by stop/clear are not reported
Notes
  • Speech is device-wide, unrelated to display areas
  • Good for welcomes, hourly chimes, alerts; watch queue backlog (limit 20) with high-frequency sends
  • Pages can also speak locally via SS.speech(text) (only on locally-hosted plugin pages, no cloud traffic — see the Plugins book)
  • The remote panel's "Sound" pane is the visual version of these commands