WMSC edge-tts proxy
GET /tts?text=...&voice=...&speed=0&pitch=0&sr=16000  -> audio/wav
GET /health                                              -> json