feat: hot-reload voice models without server restart

Voice enabled/STT model are now DB-backed (admin settings), not env
vars. Added reload_stt_model()/reload_tts_model() that clear singletons
under lock and re-trigger loading. POST /api/admin/voice/reload triggers
both in background tasks. Settings UI polls /api/voice/status every 2.5s
until models are ready, with spinner feedback.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

src/fabledassistant/services/stt.py

@@ -42,6 +42,15 @@ async def load_stt_model() -> None:
             logger.exception(_load_error)
 
 
+async def reload_stt_model() -> None:
+    """Unload the current model and reload with the current config. Safe to call at runtime."""
+    global _model, _load_error
+    async with _model_lock:
+        _model = None
+        _load_error = None
+    await load_stt_model()
+
+
 def stt_available() -> bool:
     return _model is not None
 

src/fabledassistant/services/tts.py

@@ -56,6 +56,15 @@ async def load_tts_model() -> None:
             logger.exception(_load_error)
 
 
+async def reload_tts_model() -> None:
+    """Unload and reload the TTS pipeline. Safe to call at runtime."""
+    global _pipeline, _load_error
+    async with _pipeline_lock:
+        _pipeline = None
+        _load_error = None
+    await load_tts_model()
+
+
 def tts_available() -> bool:
     return _pipeline is not None