Improve SessionProperties initialization: remove voice from args, set default for TurnDetection

examples/foundational/51-grok-realtime.py

@@ -52,7 +52,6 @@ from pipecat.runner.types import RunnerArguments
 from pipecat.runner.utils import create_transport
 from pipecat.services.grok.realtime.events import (
     SessionProperties,
-    TurnDetection,
     WebSearchTool,
     XSearchTool,
 )
@@ -172,11 +171,8 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
 
     # Configure Grok session properties
     session_properties = SessionProperties(
-        # Voice options: Ara (warm, friendly), Rex (confident), Sal (smooth),
-        # Eve (energetic), Leo (authoritative)
+        # Voice options: Ara, Rex, Sal, Eve, Leo
         voice="Ara",
-        # Enable server-side VAD for automatic turn detection
-        turn_detection=TurnDetection(type="server_vad"),
         # System instructions
         instructions="""You are a helpful and friendly AI assistant powered by Grok.
 

src/pipecat/services/grok/realtime/events.py

@@ -205,7 +205,9 @@ class SessionProperties(BaseModel):
     Parameters:
         instructions: System instructions for the assistant.
         voice: The voice the model uses to respond. Options: Ara, Rex, Sal, Eve, Leo.
-        turn_detection: Configuration for turn detection, or None for manual.
+            Defaults to "Ara".
+        turn_detection: Configuration for turn detection. Defaults to server-side VAD.
+            Set to None for manual turn detection.
         audio: Configuration for input and output audio.
         tools: Available tools for the assistant (web_search, x_search, file_search, function).
     """
@@ -215,7 +217,9 @@ class SessionProperties(BaseModel):
 
     instructions: Optional[str] = None
     voice: Optional[GrokVoice] = "Ara"
-    turn_detection: Optional[TurnDetection] = None
+    turn_detection: Optional[TurnDetection] = Field(
+        default_factory=lambda: TurnDetection(type="server_vad")
+    )
     audio: Optional[AudioConfiguration] = None
     # Tools can be ToolsSchema when provided by user, or list of dicts for API
     tools: Optional[ToolsSchema | List[GrokTool]] = None

src/pipecat/services/grok/realtime/llm.py

@@ -107,7 +107,6 @@ class GrokRealtimeLLMService(LLMService):
         self,
         *,
         api_key: str,
-        voice: events.GrokVoice = "Ara",
         base_url: str = "wss://api.x.ai/v1/realtime",
         session_properties: Optional[events.SessionProperties] = None,
         start_audio_paused: bool = False,
@@ -117,12 +116,15 @@ class GrokRealtimeLLMService(LLMService):
 
         Args:
             api_key: xAI API key for authentication.
-            voice: Voice to use for responses. Options: Ara, Rex, Sal, Eve, Leo.
-                Defaults to "Ara".
             base_url: WebSocket base URL for the realtime API.
                 Defaults to "wss://api.x.ai/v1/realtime".
             session_properties: Configuration properties for the realtime session.
-                If None, uses default SessionProperties with the specified voice.
+                If None, uses default SessionProperties with voice "Ara".
+                To set a different voice, configure it in session_properties:
+
+                    session_properties = events.SessionProperties(voice="Rex")
+
+                Available voices: Ara, Rex, Sal, Eve, Leo.
             start_audio_paused: Whether to start with audio input paused. Defaults to False.
             **kwargs: Additional arguments passed to parent LLMService.
         """
@@ -130,20 +132,11 @@ class GrokRealtimeLLMService(LLMService):
 
         self.api_key = api_key
         self.base_url = base_url
-        self._voice = voice
 
         # Initialize session_properties
-        if session_properties:
-            self._session_properties = session_properties
-            # Ensure voice is set
-            if not self._session_properties.voice:
-                self._session_properties.voice = voice
-        else:
-            self._session_properties = events.SessionProperties(
-                voice=voice,
-                turn_detection=events.TurnDetection(type="server_vad"),
-                # Audio config will be set in start() based on PipelineParams
-            )
+        self._session_properties: events.SessionProperties = (
+            session_properties or events.SessionProperties()
+        )
 
         self._audio_input_paused = start_audio_paused
         self._websocket = None