Consume InterimTranscriptionFrame and TranslationFrame in LLMUserAggregator

These frames were falling through to the else branch and being pushed
downstream, unlike TranscriptionFrame which is explicitly consumed.
This aligns with how the assistant aggregator already filters them.

src/pipecat/processors/aggregators/llm_response_universal.py

@@ -461,6 +461,10 @@ class LLMUserAggregator(LLMContextAggregator):
             await self.push_frame(frame, direction)
         elif isinstance(frame, TranscriptionFrame):
             await self._handle_transcription(frame)
+        elif isinstance(frame, (InterimTranscriptionFrame, TranslationFrame)):
+            # Interim transcriptions and translations are consumed here
+            # and not pushed downstream, same as final TranscriptionFrame.
+            pass
         elif isinstance(frame, LLMRunFrame):
             await self._handle_llm_run(frame)
         elif isinstance(frame, LLMMessagesAppendFrame):

tests/test_context_aggregators_universal.py

@@ -12,6 +12,7 @@ from pipecat.frames.frames import (
     FunctionCallFromLLM,
     FunctionCallResultFrame,
     FunctionCallsStartedFrame,
+    InterimTranscriptionFrame,
     InterruptionFrame,
     LLMContextAssistantTimestampFrame,
     LLMContextFrame,
@@ -26,6 +27,7 @@ from pipecat.frames.frames import (
     LLMThoughtTextFrame,
     StartFrame,
     TranscriptionFrame,
+    TranslationFrame,
     UserMuteStartedFrame,
     UserStartedSpeakingFrame,
     UserStoppedSpeakingFrame,
@@ -428,6 +430,44 @@ class TestLLMUserAggregator(unittest.IsolatedAsyncioTestCase):
             ignore_start=False,
         )
 
+    async def test_interim_transcription_not_pushed_downstream(self):
+        """InterimTranscriptionFrame should be consumed and not pushed downstream."""
+        context = LLMContext()
+        pipeline = Pipeline([LLMUserAggregator(context)])
+
+        frames_to_send = [
+            InterimTranscriptionFrame(text="Hel", user_id="", timestamp="now"),
+            InterimTranscriptionFrame(text="Hello", user_id="", timestamp="now"),
+        ]
+        # The interim transcription triggers a user turn start via the default
+        # TranscriptionUserTurnStartStrategy, so we expect turn-related frames
+        # but NOT the InterimTranscriptionFrame itself.
+        expected_down_frames = [
+            UserStartedSpeakingFrame,
+            InterruptionFrame,
+        ]
+        (down_frames, _) = await run_test(
+            pipeline,
+            frames_to_send=frames_to_send,
+            expected_down_frames=expected_down_frames,
+        )
+        self.assertFalse(any(isinstance(f, InterimTranscriptionFrame) for f in down_frames))
+
+    async def test_translation_not_pushed_downstream(self):
+        """TranslationFrame should be consumed and not pushed downstream."""
+        context = LLMContext()
+        pipeline = Pipeline([LLMUserAggregator(context)])
+
+        frames_to_send = [
+            TranslationFrame(text="Hola!", user_id="", timestamp="now", language="es"),
+        ]
+        # No downstream frames expected — translations are consumed.
+        await run_test(
+            pipeline,
+            frames_to_send=frames_to_send,
+            expected_down_frames=[],
+        )
+
 
 class TestLLMAssistantAggregator(unittest.IsolatedAsyncioTestCase):
     async def test_empty(self):