Fix default voice to use server configuration

Changed voice field in NotifyRequest from hardcoded default to None,
allowing the route handler to use settings.default_voice from .env.
Added EnvironmentFile directive to systemd service for reliable config loading.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

app/models.py

@@ -30,11 +30,11 @@ class NotifyRequest(BaseModel):
         ),
     ]
     voice: Annotated[
-        str,
+        str | None,
         Field(
-            default="en_US-lessac-medium",
+            default=None,
             pattern=r"^[\w-]+$",
-            description="Piper voice model name",
+            description="Piper voice model name (uses server default if not specified)",
         ),
     ]
     rate: Annotated[

voice-server.service

@@ -7,6 +7,7 @@ After=network.target sound.target
 Type=simple
 WorkingDirectory=/mnt/NV2/Development/voice-server
 Environment=PATH=/mnt/NV2/Development/voice-server/.venv/bin:/usr/bin:/bin
+EnvironmentFile=/mnt/NV2/Development/voice-server/.env
 ExecStart=/mnt/NV2/Development/voice-server/.venv/bin/uvicorn app.main:app --host 0.0.0.0 --port 8888
 Restart=on-failure
 RestartSec=5