OpenAI Realtime needs the assistant context aggregator to have expect_stripped_words=False

CHANGELOG.md

@@ -14,7 +14,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
   ```python
   context = LLMContext(messages, tools)
-  context_aggregator = LLMContextAggregatorPair(context)
+  context_aggregator = LLMContextAggregatorPair(
+    context,
+    # This part is `OpenAIRealtimeLLMService`-specific.
+    # `expect_stripped_words=False` needed when OpenAI Realtime used with
+    # "audio" modality (the default).
+    assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
+  )
   ```
 
   (Note that even though `OpenAIRealtimeLLMService` now supports the universal

examples/foundational/19-openai-realtime.py

@@ -21,6 +21,7 @@ from pipecat.pipeline.pipeline import Pipeline
 from pipecat.pipeline.runner import PipelineRunner
 from pipecat.pipeline.task import PipelineParams, PipelineTask
 from pipecat.processors.aggregators.llm_context import LLMContext
+from pipecat.processors.aggregators.llm_response import LLMAssistantAggregatorParams
 from pipecat.processors.aggregators.llm_response_universal import LLMContextAggregatorPair
 from pipecat.processors.transcript_processor import TranscriptProcessor
 from pipecat.runner.types import RunnerArguments
@@ -186,7 +187,12 @@ Remember, your responses should be short. Just one or two sentences, usually. Re
         tools,
     )
 
-    context_aggregator = LLMContextAggregatorPair(context)
+    context_aggregator = LLMContextAggregatorPair(
+        context,
+        # `expect_stripped_words=False` needed when OpenAI Realtime used with
+        # "audio" modality (the default)
+        assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
+    )
 
     pipeline = Pipeline(
         [

examples/foundational/19a-azure-realtime.py

@@ -19,6 +19,7 @@ from pipecat.pipeline.pipeline import Pipeline
 from pipecat.pipeline.runner import PipelineRunner
 from pipecat.pipeline.task import PipelineParams, PipelineTask
 from pipecat.processors.aggregators.llm_context import LLMContext
+from pipecat.processors.aggregators.llm_response import LLMAssistantAggregatorParams
 from pipecat.processors.aggregators.llm_response_universal import LLMContextAggregatorPair
 from pipecat.runner.types import RunnerArguments
 from pipecat.runner.utils import create_transport
@@ -174,7 +175,12 @@ Remember, your responses should be short. Just one or two sentences, usually. Re
         tools,
     )
 
-    context_aggregator = LLMContextAggregatorPair(context)
+    context_aggregator = LLMContextAggregatorPair(
+        context,
+        # `expect_stripped_words=False` needed when OpenAI Realtime used with
+        # "audio" modality (the default)
+        assistant_params=LLMAssistantAggregatorParams(expect_stripped_words=False),
+    )
 
     pipeline = Pipeline(
         [

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

@@ -895,6 +895,7 @@ class OpenAIRealtimeLLMService(LLMService):
         )
 
         context = LLMContext.from_openai_context(context)
+        assistant_params.expect_stripped_words = False
         return LLMContextAggregatorPair(
             context, user_params=user_params, assistant_params=assistant_params
         )