diff --git a/api/src/routers/development.py b/api/src/routers/development.py
index 82d5c40..e1cdbab 100644
--- a/api/src/routers/development.py
+++ b/api/src/routers/development.py
@@ -294,6 +294,7 @@ async def create_captioned_speech(
                 request.response_format,
                 is_first_chunk=True,
                 is_last_chunk=False,
+                trim_audio=False,
             )
             
             # Convert to requested format with proper finalization
diff --git a/api/src/routers/openai_compatible.py b/api/src/routers/openai_compatible.py
index 73b3923..c4036d6 100644
--- a/api/src/routers/openai_compatible.py
+++ b/api/src/routers/openai_compatible.py
@@ -294,6 +294,7 @@ async def create_speech(
                 request.response_format,
                 is_first_chunk=True,
                 is_last_chunk=False,
+                trim_audio=False
             )
             
             # Convert to requested format with proper finalization
diff --git a/api/src/services/audio.py b/api/src/services/audio.py
index 222b5b5..eb31c59 100644
--- a/api/src/services/audio.py
+++ b/api/src/services/audio.py
@@ -119,6 +119,7 @@ class AudioService:
         chunk_text: str = "",
         is_first_chunk: bool = True,
         is_last_chunk: bool = False,
+        trim_audio: bool = True,
         normalizer: AudioNormalizer = None,
     ) -> Tuple[AudioChunk]:
         """Convert audio data to specified format with streaming support
@@ -147,7 +148,9 @@ class AudioService:
                 normalizer = AudioNormalizer()
             
             audio_chunk.audio = normalizer.normalize(audio_chunk.audio)
-            audio_chunk = AudioService.trim_audio(audio_chunk,chunk_text,speed,is_last_chunk,normalizer)
+            
+            if trim_audio == True:
+                audio_chunk = AudioService.trim_audio(audio_chunk,chunk_text,speed,is_last_chunk,normalizer)
             
             # Get or create format-specific writer
             writer_key = f"{output_format}_{sample_rate}"
diff --git a/api/src/services/tts_service.py b/api/src/services/tts_service.py
index d2aaaf2..716c603 100644
--- a/api/src/services/tts_service.py
+++ b/api/src/services/tts_service.py
@@ -65,7 +65,7 @@ class TTSService:
                         yield AudioChunk(np.array([], dtype=np.int16),output=b'')
                         return
                     chunk_data = await AudioService.convert_audio(
-                        AudioChunk(np.array([0], dtype=np.float32)),  # Dummy data for type checking
+                        AudioChunk(np.array([], dtype=np.float32)),  # Dummy data for type checking
                         24000,
                         output_format,
                         speed,
@@ -225,6 +225,8 @@ class TTSService:
                 return voice, combined_path
             else:
                 # Single voice
+                if "(" in voice and ")" in voice:
+                    voice = voice.split("(")[0].strip()
                 path = await self._voice_manager.get_voice_path(voice)
                 if not path:
                     raise RuntimeError(f"Voice not found: {voice}")
@@ -341,8 +343,9 @@ class TTSService:
   
         try:
             async for audio_stream_data in self.generate_audio_stream(text,voice,speed=speed,normalization_options=normalization_options,return_timestamps=return_timestamps,lang_code=lang_code,output_format=None):
+                if len(audio_stream_data.audio) > 0:
+                    audio_data_chunks.append(audio_stream_data)
 
-                audio_data_chunks.append(audio_stream_data)
 
             combined_audio_data=AudioChunk.combine(audio_data_chunks)
             return combined_audio_data