diff --git a/src/pipecat/services/stt_service.py b/src/pipecat/services/stt_service.py
index 44c7af3b4..21422d6ef 100644
--- a/src/pipecat/services/stt_service.py
+++ b/src/pipecat/services/stt_service.py
@@ -497,11 +497,11 @@ class SegmentedSTTService(STTService):
         wav.close()
         content.seek(0)
 
-        await self.process_generator(self.run_stt(content.read()))
-
         # Start clean.
         self._audio_buffer.clear()
 
+        await self.process_generator(self.run_stt(content.read()))
+
     async def process_audio_frame(self, frame: AudioRawFrame, direction: FrameDirection):
         """Process audio frames by buffering them for segmented transcription.