TTS services: Add aggregate_sentences arg

CHANGELOG.md

@@ -9,6 +9,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
+- Added an `aggregate_sentences` arg in `CartesiaTTSService`,
+  `ElevenLabsTTSService`, `NeuphonicTTSService` and `RimeTTSService`, where the
+  default value is True. When `aggregate_sentences` is True, the `TTSService`
+  aggregates the LLM streamed tokens into sentences by default. Note: setting
+  the value to False requires a custom processor before the `TTSService` to
+  aggregate LLM tokens.
+
 - Added call hang-up error handling in `TwilioFrameSerializer`, which handles
   the case where the user has hung up before the `TwilioFrameSerializer` hangs
   up the call.

src/pipecat/services/cartesia/tts.py

@@ -121,6 +121,7 @@ class CartesiaTTSService(AudioContextWordTTSService):
         container: str = "raw",
         params: Optional[InputParams] = None,
         text_aggregator: Optional[BaseTextAggregator] = None,
+        aggregate_sentences: Optional[bool] = True,
         **kwargs,
     ):
         """Initialize the Cartesia TTS service.
@@ -136,6 +137,7 @@ class CartesiaTTSService(AudioContextWordTTSService):
             container: Audio container format.
             params: Additional input parameters for voice customization.
             text_aggregator: Custom text aggregator for processing input text.
+            aggregate_sentences: Whether to aggregate sentences within the TTSService.
             **kwargs: Additional arguments passed to the parent service.
         """
         # Aggregating sentences still gives cleaner-sounding results and fewer
@@ -149,7 +151,7 @@ class CartesiaTTSService(AudioContextWordTTSService):
         # can use those to generate text frames ourselves aligned with the
         # playout timing of the audio!
         super().__init__(
-            aggregate_sentences=True,
+            aggregate_sentences=aggregate_sentences,
             push_text_frames=False,
             pause_frame_processing=True,
             sample_rate=sample_rate,

src/pipecat/services/elevenlabs/tts.py

@@ -238,6 +238,7 @@ class ElevenLabsTTSService(AudioContextWordTTSService):
         url: str = "wss://api.elevenlabs.io",
         sample_rate: Optional[int] = None,
         params: Optional[InputParams] = None,
+        aggregate_sentences: Optional[bool] = True,
         **kwargs,
     ):
         """Initialize the ElevenLabs TTS service.
@@ -249,6 +250,7 @@ class ElevenLabsTTSService(AudioContextWordTTSService):
             url: WebSocket URL for ElevenLabs TTS API.
             sample_rate: Audio sample rate. If None, uses default.
             params: Additional input parameters for voice customization.
+            aggregate_sentences: Whether to aggregate sentences within the TTSService.
             **kwargs: Additional arguments passed to the parent service.
         """
         # Aggregating sentences still gives cleaner-sounding results and fewer
@@ -266,7 +268,7 @@ class ElevenLabsTTSService(AudioContextWordTTSService):
         # speaking for a while, so we want the parent class to send TTSStopFrame
         # after a short period not receiving any audio.
         super().__init__(
-            aggregate_sentences=True,
+            aggregate_sentences=aggregate_sentences,
             push_text_frames=False,
             push_stop_frames=True,
             pause_frame_processing=True,

src/pipecat/services/neuphonic/tts.py

@@ -110,6 +110,7 @@ class NeuphonicTTSService(InterruptibleTTSService):
         sample_rate: Optional[int] = 22050,
         encoding: str = "pcm_linear",
         params: Optional[InputParams] = None,
+        aggregate_sentences: Optional[bool] = True,
         **kwargs,
     ):
         """Initialize the Neuphonic TTS service.
@@ -121,10 +122,11 @@ class NeuphonicTTSService(InterruptibleTTSService):
             sample_rate: Audio sample rate in Hz. Defaults to 22050.
             encoding: Audio encoding format. Defaults to "pcm_linear".
             params: Additional input parameters for TTS configuration.
+            aggregate_sentences: Whether to aggregate sentences within the TTSService.
             **kwargs: Additional arguments passed to parent InterruptibleTTSService.
         """
         super().__init__(
-            aggregate_sentences=True,
+            aggregate_sentences=aggregate_sentences,
             push_text_frames=False,
             push_stop_frames=True,
             stop_frame_timeout_s=2.0,

src/pipecat/services/rime/tts.py

@@ -99,6 +99,7 @@ class RimeTTSService(AudioContextWordTTSService):
         sample_rate: Optional[int] = None,
         params: Optional[InputParams] = None,
         text_aggregator: Optional[BaseTextAggregator] = None,
+        aggregate_sentences: Optional[bool] = True,
         **kwargs,
     ):
         """Initialize Rime TTS service.
@@ -111,11 +112,12 @@ class RimeTTSService(AudioContextWordTTSService):
             sample_rate: Audio sample rate in Hz.
             params: Additional configuration parameters.
             text_aggregator: Custom text aggregator for processing input text.
+            aggregate_sentences: Whether to aggregate sentences within the TTSService.
             **kwargs: Additional arguments passed to parent class.
         """
         # Initialize with parent class settings for proper frame handling
         super().__init__(
-            aggregate_sentences=True,
+            aggregate_sentences=aggregate_sentences,
             push_text_frames=False,
             push_stop_frames=True,
             pause_frame_processing=True,