chore: minor fixes

api/services/configuration/registry.py

@@ -122,7 +122,7 @@ GROQ_MODELS = [
     "openai/gpt-oss-120b",
 ]
 AZURE_MODELS = ["gpt-4.1-mini"]
-DOGRAH_LLM_MODELS = ["default", "accurate", "fast", "lite", "zen", "zen_lite"]
+DOGRAH_LLM_MODELS = ["default", "accurate", "fast", "lite", "zen"]
 
 
 @register_llm
@@ -266,9 +266,13 @@ SARVAM_LANGUAGES = [
 @register_tts
 class SarvamTTSConfiguration(BaseTTSConfiguration):
     provider: Literal[ServiceProviders.SARVAM] = ServiceProviders.SARVAM
-    model: str = Field(default="bulbul:v2", json_schema_extra={"examples": SARVAM_TTS_MODELS})
+    model: str = Field(
+        default="bulbul:v2", json_schema_extra={"examples": SARVAM_TTS_MODELS}
+    )
     voice: str = Field(default="anushka", json_schema_extra={"examples": SARVAM_VOICES})
-    language: str = Field(default="hi-IN", json_schema_extra={"examples": SARVAM_LANGUAGES})
+    language: str = Field(
+        default="hi-IN", json_schema_extra={"examples": SARVAM_LANGUAGES}
+    )
     api_key: str
 
 
@@ -372,8 +376,12 @@ SARVAM_STT_MODELS = ["saarika:v2.5", "saaras:v2"]
 @register_stt
 class SarvamSTTConfiguration(BaseSTTConfiguration):
     provider: Literal[ServiceProviders.SARVAM] = ServiceProviders.SARVAM
-    model: str = Field(default="saarika:v2.5", json_schema_extra={"examples": SARVAM_STT_MODELS})
+    model: str = Field(
-    language: str = Field(default="hi-IN", json_schema_extra={"examples": SARVAM_LANGUAGES})
+        default="saarika:v2.5", json_schema_extra={"examples": SARVAM_STT_MODELS}
+    )
+    language: str = Field(
+        default="hi-IN", json_schema_extra={"examples": SARVAM_LANGUAGES}
+    )
     api_key: str
 
 

api/services/looptalk/core/pipeline_builder.py

@@ -91,7 +91,6 @@ class LoopTalkPipelineBuilder:
         # Create engine first (needed for create_pipeline_components)
         engine = PipecatEngine(
             llm=llm,
-            tts=tts,
             workflow=workflow_graph,
             call_context_vars={},
             workflow_run_id=None,  # LoopTalk doesn't have workflow runs

api/services/pipecat/run_pipeline.py

@@ -467,7 +467,6 @@ async def _run_pipeline(
 
     engine = PipecatEngine(
         llm=llm,
-        tts=tts,
         workflow=workflow_graph,
         call_context_vars=merged_call_context_vars,
         workflow_run_id=workflow_run_id,

api/services/pipecat/service_factory.py

@@ -31,6 +31,9 @@ if TYPE_CHECKING:
 
 def create_stt_service(user_config):
     """Create and return appropriate STT service based on user configuration"""
+    logger.info(
+        f"Creating STT service: provider={user_config.stt.provider}, model={user_config.stt.model}"
+    )
     if user_config.stt.provider == ServiceProviders.DEEPGRAM.value:
         # Use language from user config, defaulting to "multi" for multilingual support
         language = getattr(user_config.stt, "language", None) or "multi"
@@ -110,6 +113,9 @@ def create_tts_service(user_config, audio_config: "AudioConfig"):
         user_config: User configuration containing TTS settings
         transport_type: Type of transport (e.g., 'stasis', 'twilio', 'webrtc')
     """
+    logger.info(
+        f"Creating TTS service: provider={user_config.tts.provider}, model={user_config.tts.model}"
+    )
     # Create function call filter to prevent TTS from speaking function call tags
     xml_function_tag_filter = XMLFunctionTagFilter()
     if user_config.tts.provider == ServiceProviders.DEEPGRAM.value:
@@ -186,6 +192,9 @@ def create_tts_service(user_config, audio_config: "AudioConfig"):
 def create_llm_service(user_config):
     """Create and return appropriate LLM service based on user configuration"""
     model = user_config.llm.model
+    logger.info(
+        f"Creating LLM service: provider={user_config.llm.provider}, model={model}"
+    )
     if user_config.llm.provider == ServiceProviders.OPENAI.value:
         if "gpt-5" in model:
             return OpenAILLMService(
@@ -251,6 +260,6 @@ def create_voicemail_classification_llm():
 
     return OpenAILLMService(
         api_key=api_key,
-        model="gpt-4o-mini",
+        model="gpt-4o",
         params=OpenAILLMService.InputParams(temperature=0.0),
     )

api/services/workflow/pipecat_engine.py

@@ -1,4 +1,4 @@
-from typing import TYPE_CHECKING, Any, Awaitable, Callable, Optional, Union
+from typing import TYPE_CHECKING, Awaitable, Callable, Optional, Union
 
 from api.services.workflow.disposition_mapper import (
     apply_disposition_mapping,
@@ -57,7 +57,6 @@ class PipecatEngine:
         task: Optional[PipelineTask] = None,
         llm: Optional["LLMService"] = None,
         context: Optional[LLMContext] = None,
-        tts: Optional[Any] = None,
         transport: Optional[BaseTransport] = None,
         workflow: WorkflowGraph,
         call_context_vars: dict,
@@ -67,7 +66,6 @@ class PipecatEngine:
         self.task = task
         self.llm = llm
         self.context = context
-        self.tts = tts
         self.transport = transport
         self.workflow = workflow
         self._call_context_vars = call_context_vars

api/services/workflow/pipecat_engine_callbacks.py

@@ -75,7 +75,7 @@ def create_user_idle_callback(engine: "PipecatEngine"):
 
         message = {
             "role": "system",
-            "content": "The user has been quiet. We will be disconnecting the call now. Wish them a good day.",
+            "content": "The user has been quiet. We will be disconnecting the call now. Wish them a good day in the language that the user has been speaking so far.",
         }
         await user_idle.push_frame(LLMMessagesAppendFrame([message], run_llm=True))
         await engine.send_end_task_frame(