sarvamai 0.1.21__tar.gz → 0.1.22a2__tar.gz

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sarvamai
-Version: 0.1.21
+Version: 0.1.22a2
 Summary: 
 Requires-Python: >=3.8,<4.0
 Classifier: Intended Audience :: Developers

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/pyproject.toml

@@ -3,7 +3,7 @@ name = "sarvamai"
 
 [tool.poetry]
 name = "sarvamai"
-version = "0.1.21"
+version = "0.1.22a2"
 description = ""
 readme = "README.md"
 authors = []

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/__init__.py

@@ -56,6 +56,7 @@ from .types import (
     NumeralsFormat,
     PingSignal,
     ReasoningEffort,
+    ResponseSpeechState,
     ResponseType,
     Role,
     SarvamModelIds,
@@ -297,6 +298,7 @@ __all__ = [
     "PingSignal",
     "PingSignalParams",
     "ReasoningEffort",
+    "ResponseSpeechState",
     "ResponseType",
     "Role",
     "SarvamAI",

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/core/client_wrapper.py

@@ -23,10 +23,10 @@ class BaseClientWrapper:
 
     def get_headers(self) -> typing.Dict[str, str]:
         headers: typing.Dict[str, str] = {
-            "User-Agent": "sarvamai/0.1.21",
+            "User-Agent": "sarvamai/0.1.22a2",
             "X-Fern-Language": "Python",
             "X-Fern-SDK-Name": "sarvamai",
-            "X-Fern-SDK-Version": "0.1.21",
+            "X-Fern-SDK-Version": "0.1.22a2",
             **(self.get_custom_headers() or {}),
         }
         headers["api-subscription-key"] = self.api_subscription_key

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/requests/speech_to_text_transcription_data.py

@@ -3,6 +3,7 @@
 import typing
 
 import typing_extensions
+from ..types.response_speech_state import ResponseSpeechState
 from .transcription_metrics import TranscriptionMetricsParams
 
 
@@ -32,4 +33,9 @@ class SpeechToTextTranscriptionDataParams(typing_extensions.TypedDict):
     BCP-47 code of detected language
     """
 
+    response_speech_state: typing_extensions.NotRequired[ResponseSpeechState]
+    """
+    Current state of speech detection and processing
+    """
+
     metrics: TranscriptionMetricsParams

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/requests/speech_to_text_translate_transcription_data.py

@@ -1,6 +1,7 @@
 # This file was auto-generated by Fern from our API Definition.
 
 import typing_extensions
+from ..types.response_speech_state import ResponseSpeechState
 from .transcription_metrics import TranscriptionMetricsParams
 
 
@@ -20,4 +21,9 @@ class SpeechToTextTranslateTranscriptionDataParams(typing_extensions.TypedDict):
     BCP-47 code of detected source language (null when language detection is in progress)
     """
 
+    response_speech_state: typing_extensions.NotRequired[ResponseSpeechState]
+    """
+    Current state of speech detection and processing
+    """
+
     metrics: TranscriptionMetricsParams

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/speech_to_text_streaming/client.py

@@ -49,11 +49,17 @@ class SpeechToTextStreamingClient:
         high_vad_sensitivity: typing.Optional[SpeechToTextStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextStreamingVadSignals] = None,
         flush_signal: typing.Optional[SpeechToTextStreamingFlushSignal] = None,
+        stream_ongoing_speech_results: typing.Optional[str] = None,
+        streaming_ongoing_requests_frame_size: typing.Optional[str] = None,
         api_subscription_key: typing.Optional[str] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.Iterator[SpeechToTextStreamingSocketClient]:
         """
-        WebSocket channel for real-time speech to text streaming
+        WebSocket channel for real-time speech to text streaming.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -78,6 +84,12 @@ class SpeechToTextStreamingClient:
         flush_signal : typing.Optional[SpeechToTextStreamingFlushSignal]
             Signal to flush the audio buffer and finalize transcription
 
+        stream_ongoing_speech_results : typing.Optional[str]
+            Enable streaming of ongoing speech results during active speech
+
+        streaming_ongoing_requests_frame_size : typing.Optional[str]
+            Frame size for streaming ongoing speech results (1-100)
+
         api_subscription_key : typing.Optional[str]
             API subscription key for authentication
 
@@ -104,6 +116,12 @@ class SpeechToTextStreamingClient:
             query_params = query_params.add("vad_signals", vad_signals)
         if flush_signal is not None:
             query_params = query_params.add("flush_signal", flush_signal)
+        if stream_ongoing_speech_results is not None:
+            query_params = query_params.add("stream_ongoing_speech_results", stream_ongoing_speech_results)
+        if streaming_ongoing_requests_frame_size is not None:
+            query_params = query_params.add(
+                "streaming_ongoing_requests_frame_size", streaming_ongoing_requests_frame_size
+            )
         ws_url = ws_url + f"?{query_params}"
         headers = self._raw_client._client_wrapper.get_headers()
         if api_subscription_key is not None:
@@ -154,11 +172,17 @@ class AsyncSpeechToTextStreamingClient:
         high_vad_sensitivity: typing.Optional[SpeechToTextStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextStreamingVadSignals] = None,
         flush_signal: typing.Optional[SpeechToTextStreamingFlushSignal] = None,
+        stream_ongoing_speech_results: typing.Optional[str] = None,
+        streaming_ongoing_requests_frame_size: typing.Optional[str] = None,
         api_subscription_key: typing.Optional[str] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.AsyncIterator[AsyncSpeechToTextStreamingSocketClient]:
         """
-        WebSocket channel for real-time speech to text streaming
+        WebSocket channel for real-time speech to text streaming.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -183,6 +207,12 @@ class AsyncSpeechToTextStreamingClient:
         flush_signal : typing.Optional[SpeechToTextStreamingFlushSignal]
             Signal to flush the audio buffer and finalize transcription
 
+        stream_ongoing_speech_results : typing.Optional[str]
+            Enable streaming of ongoing speech results during active speech
+
+        streaming_ongoing_requests_frame_size : typing.Optional[str]
+            Frame size for streaming ongoing speech results (1-100)
+
         api_subscription_key : typing.Optional[str]
             API subscription key for authentication
 
@@ -209,6 +239,12 @@ class AsyncSpeechToTextStreamingClient:
             query_params = query_params.add("vad_signals", vad_signals)
         if flush_signal is not None:
             query_params = query_params.add("flush_signal", flush_signal)
+        if stream_ongoing_speech_results is not None:
+            query_params = query_params.add("stream_ongoing_speech_results", stream_ongoing_speech_results)
+        if streaming_ongoing_requests_frame_size is not None:
+            query_params = query_params.add(
+                "streaming_ongoing_requests_frame_size", streaming_ongoing_requests_frame_size
+            )
         ws_url = ws_url + f"?{query_params}"
         headers = self._raw_client._client_wrapper.get_headers()
         if api_subscription_key is not None:

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/speech_to_text_streaming/raw_client.py

@@ -37,11 +37,17 @@ class RawSpeechToTextStreamingClient:
         high_vad_sensitivity: typing.Optional[SpeechToTextStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextStreamingVadSignals] = None,
         flush_signal: typing.Optional[SpeechToTextStreamingFlushSignal] = None,
+        stream_ongoing_speech_results: typing.Optional[str] = None,
+        streaming_ongoing_requests_frame_size: typing.Optional[str] = None,
         api_subscription_key: typing.Optional[str] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.Iterator[SpeechToTextStreamingSocketClient]:
         """
-        WebSocket channel for real-time speech to text streaming
+        WebSocket channel for real-time speech to text streaming.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -66,6 +72,12 @@ class RawSpeechToTextStreamingClient:
         flush_signal : typing.Optional[SpeechToTextStreamingFlushSignal]
             Signal to flush the audio buffer and finalize transcription
 
+        stream_ongoing_speech_results : typing.Optional[str]
+            Enable streaming of ongoing speech results during active speech
+
+        streaming_ongoing_requests_frame_size : typing.Optional[str]
+            Frame size for streaming ongoing speech results (1-100)
+
         api_subscription_key : typing.Optional[str]
             API subscription key for authentication
 
@@ -92,6 +104,12 @@ class RawSpeechToTextStreamingClient:
             query_params = query_params.add("vad_signals", vad_signals)
         if flush_signal is not None:
             query_params = query_params.add("flush_signal", flush_signal)
+        if stream_ongoing_speech_results is not None:
+            query_params = query_params.add("stream_ongoing_speech_results", stream_ongoing_speech_results)
+        if streaming_ongoing_requests_frame_size is not None:
+            query_params = query_params.add(
+                "streaming_ongoing_requests_frame_size", streaming_ongoing_requests_frame_size
+            )
         ws_url = ws_url + f"?{query_params}"
         headers = self._client_wrapper.get_headers()
         if api_subscription_key is not None:
@@ -131,11 +149,17 @@ class AsyncRawSpeechToTextStreamingClient:
         high_vad_sensitivity: typing.Optional[SpeechToTextStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextStreamingVadSignals] = None,
         flush_signal: typing.Optional[SpeechToTextStreamingFlushSignal] = None,
+        stream_ongoing_speech_results: typing.Optional[str] = None,
+        streaming_ongoing_requests_frame_size: typing.Optional[str] = None,
         api_subscription_key: typing.Optional[str] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.AsyncIterator[AsyncSpeechToTextStreamingSocketClient]:
         """
-        WebSocket channel for real-time speech to text streaming
+        WebSocket channel for real-time speech to text streaming.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -160,6 +184,12 @@ class AsyncRawSpeechToTextStreamingClient:
         flush_signal : typing.Optional[SpeechToTextStreamingFlushSignal]
             Signal to flush the audio buffer and finalize transcription
 
+        stream_ongoing_speech_results : typing.Optional[str]
+            Enable streaming of ongoing speech results during active speech
+
+        streaming_ongoing_requests_frame_size : typing.Optional[str]
+            Frame size for streaming ongoing speech results (1-100)
+
         api_subscription_key : typing.Optional[str]
             API subscription key for authentication
 
@@ -186,6 +216,12 @@ class AsyncRawSpeechToTextStreamingClient:
             query_params = query_params.add("vad_signals", vad_signals)
         if flush_signal is not None:
             query_params = query_params.add("flush_signal", flush_signal)
+        if stream_ongoing_speech_results is not None:
+            query_params = query_params.add("stream_ongoing_speech_results", stream_ongoing_speech_results)
+        if streaming_ongoing_requests_frame_size is not None:
+            query_params = query_params.add(
+                "streaming_ongoing_requests_frame_size", streaming_ongoing_requests_frame_size
+            )
         ws_url = ws_url + f"?{query_params}"
         headers = self._client_wrapper.get_headers()
         if api_subscription_key is not None:

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/speech_to_text_translate_streaming/client.py

@@ -49,11 +49,17 @@ class SpeechToTextTranslateStreamingClient:
         high_vad_sensitivity: typing.Optional[SpeechToTextTranslateStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextTranslateStreamingVadSignals] = None,
         flush_signal: typing.Optional[SpeechToTextTranslateStreamingFlushSignal] = None,
+        stream_ongoing_speech_results: typing.Optional[str] = None,
+        streaming_ongoing_requests_frame_size: typing.Optional[str] = None,
         api_subscription_key: typing.Optional[str] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.Iterator[SpeechToTextTranslateStreamingSocketClient]:
         """
-        WebSocket channel for real-time speech to text streaming with English translation
+        WebSocket channel for real-time speech to text streaming with English translation.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -75,6 +81,12 @@ class SpeechToTextTranslateStreamingClient:
         flush_signal : typing.Optional[SpeechToTextTranslateStreamingFlushSignal]
             Signal to flush the audio buffer and finalize transcription and translation
 
+        stream_ongoing_speech_results : typing.Optional[str]
+            Enable streaming of ongoing speech results during active speech
+
+        streaming_ongoing_requests_frame_size : typing.Optional[str]
+            Frame size for streaming ongoing speech results (1-100)
+
         api_subscription_key : typing.Optional[str]
             API subscription key for authentication
 
@@ -99,6 +111,12 @@ class SpeechToTextTranslateStreamingClient:
             query_params = query_params.add("vad_signals", vad_signals)
         if flush_signal is not None:
             query_params = query_params.add("flush_signal", flush_signal)
+        if stream_ongoing_speech_results is not None:
+            query_params = query_params.add("stream_ongoing_speech_results", stream_ongoing_speech_results)
+        if streaming_ongoing_requests_frame_size is not None:
+            query_params = query_params.add(
+                "streaming_ongoing_requests_frame_size", streaming_ongoing_requests_frame_size
+            )
         ws_url = ws_url + f"?{query_params}"
         headers = self._raw_client._client_wrapper.get_headers()
         if api_subscription_key is not None:
@@ -148,11 +166,17 @@ class AsyncSpeechToTextTranslateStreamingClient:
         high_vad_sensitivity: typing.Optional[SpeechToTextTranslateStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextTranslateStreamingVadSignals] = None,
         flush_signal: typing.Optional[SpeechToTextTranslateStreamingFlushSignal] = None,
+        stream_ongoing_speech_results: typing.Optional[str] = None,
+        streaming_ongoing_requests_frame_size: typing.Optional[str] = None,
         api_subscription_key: typing.Optional[str] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.AsyncIterator[AsyncSpeechToTextTranslateStreamingSocketClient]:
         """
-        WebSocket channel for real-time speech to text streaming with English translation
+        WebSocket channel for real-time speech to text streaming with English translation.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -174,6 +198,12 @@ class AsyncSpeechToTextTranslateStreamingClient:
         flush_signal : typing.Optional[SpeechToTextTranslateStreamingFlushSignal]
             Signal to flush the audio buffer and finalize transcription and translation
 
+        stream_ongoing_speech_results : typing.Optional[str]
+            Enable streaming of ongoing speech results during active speech
+
+        streaming_ongoing_requests_frame_size : typing.Optional[str]
+            Frame size for streaming ongoing speech results (1-100)
+
         api_subscription_key : typing.Optional[str]
             API subscription key for authentication
 
@@ -198,6 +228,12 @@ class AsyncSpeechToTextTranslateStreamingClient:
             query_params = query_params.add("vad_signals", vad_signals)
         if flush_signal is not None:
             query_params = query_params.add("flush_signal", flush_signal)
+        if stream_ongoing_speech_results is not None:
+            query_params = query_params.add("stream_ongoing_speech_results", stream_ongoing_speech_results)
+        if streaming_ongoing_requests_frame_size is not None:
+            query_params = query_params.add(
+                "streaming_ongoing_requests_frame_size", streaming_ongoing_requests_frame_size
+            )
         ws_url = ws_url + f"?{query_params}"
         headers = self._raw_client._client_wrapper.get_headers()
         if api_subscription_key is not None:

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/speech_to_text_translate_streaming/raw_client.py

@@ -37,11 +37,17 @@ class RawSpeechToTextTranslateStreamingClient:
         high_vad_sensitivity: typing.Optional[SpeechToTextTranslateStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextTranslateStreamingVadSignals] = None,
         flush_signal: typing.Optional[SpeechToTextTranslateStreamingFlushSignal] = None,
+        stream_ongoing_speech_results: typing.Optional[str] = None,
+        streaming_ongoing_requests_frame_size: typing.Optional[str] = None,
         api_subscription_key: typing.Optional[str] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.Iterator[SpeechToTextTranslateStreamingSocketClient]:
         """
-        WebSocket channel for real-time speech to text streaming with English translation
+        WebSocket channel for real-time speech to text streaming with English translation.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -63,6 +69,12 @@ class RawSpeechToTextTranslateStreamingClient:
         flush_signal : typing.Optional[SpeechToTextTranslateStreamingFlushSignal]
             Signal to flush the audio buffer and finalize transcription and translation
 
+        stream_ongoing_speech_results : typing.Optional[str]
+            Enable streaming of ongoing speech results during active speech
+
+        streaming_ongoing_requests_frame_size : typing.Optional[str]
+            Frame size for streaming ongoing speech results (1-100)
+
         api_subscription_key : typing.Optional[str]
             API subscription key for authentication
 
@@ -87,6 +99,12 @@ class RawSpeechToTextTranslateStreamingClient:
             query_params = query_params.add("vad_signals", vad_signals)
         if flush_signal is not None:
             query_params = query_params.add("flush_signal", flush_signal)
+        if stream_ongoing_speech_results is not None:
+            query_params = query_params.add("stream_ongoing_speech_results", stream_ongoing_speech_results)
+        if streaming_ongoing_requests_frame_size is not None:
+            query_params = query_params.add(
+                "streaming_ongoing_requests_frame_size", streaming_ongoing_requests_frame_size
+            )
         ws_url = ws_url + f"?{query_params}"
         headers = self._client_wrapper.get_headers()
         if api_subscription_key is not None:
@@ -125,11 +143,17 @@ class AsyncRawSpeechToTextTranslateStreamingClient:
         high_vad_sensitivity: typing.Optional[SpeechToTextTranslateStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextTranslateStreamingVadSignals] = None,
         flush_signal: typing.Optional[SpeechToTextTranslateStreamingFlushSignal] = None,
+        stream_ongoing_speech_results: typing.Optional[str] = None,
+        streaming_ongoing_requests_frame_size: typing.Optional[str] = None,
         api_subscription_key: typing.Optional[str] = None,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.AsyncIterator[AsyncSpeechToTextTranslateStreamingSocketClient]:
         """
-        WebSocket channel for real-time speech to text streaming with English translation
+        WebSocket channel for real-time speech to text streaming with English translation.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -151,6 +175,12 @@ class AsyncRawSpeechToTextTranslateStreamingClient:
         flush_signal : typing.Optional[SpeechToTextTranslateStreamingFlushSignal]
             Signal to flush the audio buffer and finalize transcription and translation
 
+        stream_ongoing_speech_results : typing.Optional[str]
+            Enable streaming of ongoing speech results during active speech
+
+        streaming_ongoing_requests_frame_size : typing.Optional[str]
+            Frame size for streaming ongoing speech results (1-100)
+
         api_subscription_key : typing.Optional[str]
             API subscription key for authentication
 
@@ -175,6 +205,12 @@ class AsyncRawSpeechToTextTranslateStreamingClient:
             query_params = query_params.add("vad_signals", vad_signals)
         if flush_signal is not None:
             query_params = query_params.add("flush_signal", flush_signal)
+        if stream_ongoing_speech_results is not None:
+            query_params = query_params.add("stream_ongoing_speech_results", stream_ongoing_speech_results)
+        if streaming_ongoing_requests_frame_size is not None:
+            query_params = query_params.add(
+                "streaming_ongoing_requests_frame_size", streaming_ongoing_requests_frame_size
+            )
         ws_url = ws_url + f"?{query_params}"
         headers = self._client_wrapper.get_headers()
         if api_subscription_key is not None:

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/text_to_speech_streaming/client.py

@@ -44,8 +44,11 @@ class TextToSpeechStreamingClient:
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.Iterator[TextToSpeechStreamingSocketClient]:
         """
-        Bidirectional WebSocket channel for real-time TTS synthesis.
-        Supports streaming, flushing, config updates, and audio playback.
+        WebSocket channel for real-time TTS synthesis.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -120,8 +123,11 @@ class AsyncTextToSpeechStreamingClient:
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.AsyncIterator[AsyncTextToSpeechStreamingSocketClient]:
         """
-        Bidirectional WebSocket channel for real-time TTS synthesis.
-        Supports streaming, flushing, config updates, and audio playback.
+        WebSocket channel for real-time TTS synthesis.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/text_to_speech_streaming/raw_client.py

@@ -32,8 +32,11 @@ class RawTextToSpeechStreamingClient:
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.Iterator[TextToSpeechStreamingSocketClient]:
         """
-        Bidirectional WebSocket channel for real-time TTS synthesis.
-        Supports streaming, flushing, config updates, and audio playback.
+        WebSocket channel for real-time TTS synthesis.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------
@@ -97,8 +100,11 @@ class AsyncRawTextToSpeechStreamingClient:
         request_options: typing.Optional[RequestOptions] = None,
     ) -> typing.AsyncIterator[AsyncTextToSpeechStreamingSocketClient]:
         """
-        Bidirectional WebSocket channel for real-time TTS synthesis.
-        Supports streaming, flushing, config updates, and audio playback.
+        WebSocket channel for real-time TTS synthesis.
+
+        **Note:** This API Reference page is provided for informational purposes only.
+        The Try It playground may not provide the best experience for streaming audio.
+        For optimal streaming performance, please use the SDK or implement your own WebSocket client.
 
         Parameters
         ----------

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/types/__init__.py

@@ -57,6 +57,7 @@ from .language_identification_response import LanguageIdentificationResponse
 from .numerals_format import NumeralsFormat
 from .ping_signal import PingSignal
 from .reasoning_effort import ReasoningEffort
+from .response_speech_state import ResponseSpeechState
 from .response_type import ResponseType
 from .role import Role
 from .sarvam_model_ids import SarvamModelIds
@@ -156,6 +157,7 @@ __all__ = [
     "NumeralsFormat",
     "PingSignal",
     "ReasoningEffort",
+    "ResponseSpeechState",
     "ResponseType",
     "Role",
     "SarvamModelIds",

sarvamai-0.1.22a2/src/sarvamai/types/completion_event_flag.py

@@ -0,0 +1,3 @@
+# This file was auto-generated by Fern from our API Definition.
+
+CompletionEventFlag = bool

sarvamai-0.1.22a2/src/sarvamai/types/response_speech_state.py

@@ -0,0 +1,7 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing
+
+ResponseSpeechState = typing.Union[
+    typing.Literal["START_SPEECH_TENTATIVE", "VAD_MISFIRE", "START_SPEECH", "END_SPEECH", "ONGOING"], typing.Any
+]

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/types/speech_to_text_transcription_data.py

@@ -4,6 +4,7 @@ import typing
 
 import pydantic
 from ..core.pydantic_utilities import IS_PYDANTIC_V2, UniversalBaseModel
+from .response_speech_state import ResponseSpeechState
 from .transcription_metrics import TranscriptionMetrics
 
 
@@ -33,6 +34,11 @@ class SpeechToTextTranscriptionData(UniversalBaseModel):
     BCP-47 code of detected language
     """
 
+    response_speech_state: typing.Optional[ResponseSpeechState] = pydantic.Field(default=None)
+    """
+    Current state of speech detection and processing
+    """
+
     metrics: TranscriptionMetrics
 
     if IS_PYDANTIC_V2:

{sarvamai-0.1.21 → sarvamai-0.1.22a2}/src/sarvamai/types/speech_to_text_translate_transcription_data.py

@@ -4,6 +4,7 @@ import typing
 
 import pydantic
 from ..core.pydantic_utilities import IS_PYDANTIC_V2, UniversalBaseModel
+from .response_speech_state import ResponseSpeechState
 from .transcription_metrics import TranscriptionMetrics
 
 
@@ -23,6 +24,11 @@ class SpeechToTextTranslateTranscriptionData(UniversalBaseModel):
     BCP-47 code of detected source language (null when language detection is in progress)
     """
 
+    response_speech_state: typing.Optional[ResponseSpeechState] = pydantic.Field(default=None)
+    """
+    Current state of speech detection and processing
+    """
+
     metrics: TranscriptionMetrics
 
     if IS_PYDANTIC_V2:

sarvamai-0.1.21/src/sarvamai/types/completion_event_flag.py

@@ -1,5 +0,0 @@
-# This file was auto-generated by Fern from our API Definition.
-
-import typing
-
-CompletionEventFlag = typing.Union[typing.Literal["true", "false"], typing.Any]