# This file was auto-generated by Fern from our API Definition. import contextlib import json import typing from json.decoder import JSONDecodeError from ..core.api_error import ApiError from ..core.client_wrapper import AsyncClientWrapper, SyncClientWrapper from ..core.http_response import AsyncHttpResponse, HttpResponse from ..core.jsonable_encoder import jsonable_encoder from ..core.request_options import RequestOptions from ..core.serialization import convert_and_respect_annotation_metadata from ..core.unchecked_base_model import construct_type from ..errors.unprocessable_entity_error import UnprocessableEntityError from ..types.audio_with_timestamps_response import AudioWithTimestampsResponse from ..types.http_validation_error import HttpValidationError from ..types.pronunciation_dictionary_version_locator import PronunciationDictionaryVersionLocator from ..types.streaming_audio_chunk_with_timestamps_response import StreamingAudioChunkWithTimestampsResponse from ..types.voice_settings import VoiceSettings from .types.body_text_to_speech_full_apply_text_normalization import BodyTextToSpeechFullApplyTextNormalization from .types.body_text_to_speech_full_with_timestamps_apply_text_normalization import ( BodyTextToSpeechFullWithTimestampsApplyTextNormalization, ) from .types.body_text_to_speech_stream_apply_text_normalization import BodyTextToSpeechStreamApplyTextNormalization from .types.body_text_to_speech_stream_with_timestamps_apply_text_normalization import ( BodyTextToSpeechStreamWithTimestampsApplyTextNormalization, ) from .types.text_to_speech_convert_request_output_format import TextToSpeechConvertRequestOutputFormat from .types.text_to_speech_convert_with_timestamps_request_output_format import ( TextToSpeechConvertWithTimestampsRequestOutputFormat, ) from .types.text_to_speech_stream_request_output_format import TextToSpeechStreamRequestOutputFormat from .types.text_to_speech_stream_with_timestamps_request_output_format import ( TextToSpeechStreamWithTimestampsRequestOutputFormat, ) # this is used as the default value for optional parameters OMIT = typing.cast(typing.Any, ...) class RawTextToSpeechClient: def __init__(self, *, client_wrapper: SyncClientWrapper): self._client_wrapper = client_wrapper @contextlib.contextmanager def convert( self, voice_id: str, *, text: str, enable_logging: typing.Optional[bool] = None, optimize_streaming_latency: typing.Optional[int] = None, output_format: typing.Optional[TextToSpeechConvertRequestOutputFormat] = None, model_id: typing.Optional[str] = OMIT, language_code: typing.Optional[str] = OMIT, voice_settings: typing.Optional[VoiceSettings] = OMIT, pronunciation_dictionary_locators: typing.Optional[ typing.Sequence[PronunciationDictionaryVersionLocator] ] = OMIT, seed: typing.Optional[int] = OMIT, previous_text: typing.Optional[str] = OMIT, next_text: typing.Optional[str] = OMIT, previous_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, next_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, use_pvc_as_ivc: typing.Optional[bool] = OMIT, apply_text_normalization: typing.Optional[BodyTextToSpeechFullApplyTextNormalization] = OMIT, apply_language_text_normalization: typing.Optional[bool] = OMIT, request_options: typing.Optional[RequestOptions] = None, ) -> typing.Iterator[HttpResponse[typing.Iterator[bytes]]]: """ Converts text into speech using a voice of your choice and returns audio. Parameters ---------- voice_id : str ID of the voice to be used. Use the [Get voices](/docs/api-reference/voices/search) endpoint list all the available voices. text : str The text that will get converted into speech. enable_logging : typing.Optional[bool] When enable_logging is set to false zero retention mode will be used for the request. This will mean history features are unavailable for this request, including request stitching. Zero retention mode may only be used by enterprise customers. optimize_streaming_latency : typing.Optional[int] You can turn on latency optimizations at some cost of quality. The best possible final latency varies by model. Possible values: 0 - default mode (no latency optimizations) 1 - normal latency optimizations (about 50% of possible latency improvement of option 3) 2 - strong latency optimizations (about 75% of possible latency improvement of option 3) 3 - max latency optimizations 4 - max latency optimizations, but also with text normalizer turned off for even more latency savings (best latency, but can mispronounce eg numbers and dates). Defaults to None. output_format : typing.Optional[TextToSpeechConvertRequestOutputFormat] Output format of the generated audio. Formatted as codec_sample_rate_bitrate. So an mp3 with 22.05kHz sample rate at 32kbs is represented as mp3_22050_32. MP3 with 192kbps bitrate requires you to be subscribed to Creator tier or above. PCM with 44.1kHz sample rate requires you to be subscribed to Pro tier or above. Note that the μ-law format (sometimes written mu-law, often approximated as u-law) is commonly used for Twilio audio inputs. model_id : typing.Optional[str] Identifier of the model that will be used, you can query them using GET /v1/models. The model needs to have support for text to speech, you can check this using the can_do_text_to_speech property. language_code : typing.Optional[str] Language code (ISO 639-1) used to enforce a language for the model and text normalization. If the model does not support provided language code, an error will be returned. voice_settings : typing.Optional[VoiceSettings] Voice settings overriding stored settings for the given voice. They are applied only on the given request. pronunciation_dictionary_locators : typing.Optional[typing.Sequence[PronunciationDictionaryVersionLocator]] A list of pronunciation dictionary locators (id, version_id) to be applied to the text. They will be applied in order. You may have up to 3 locators per request seed : typing.Optional[int] If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed. Must be integer between 0 and 4294967295. previous_text : typing.Optional[str] The text that came before the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. next_text : typing.Optional[str] The text that comes after the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. previous_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that were generated before this generation. Can be used to improve the speech's continuity when splitting up a large task into multiple requests. The results will be best when the same model is used across the generations. In case both previous_text and previous_request_ids is send, previous_text will be ignored. A maximum of 3 request_ids can be send. next_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that come after this generation. next_request_ids is especially useful for maintaining the speech's continuity when regenerating a sample that has had some audio quality issues. For example, if you have generated 3 speech clips, and you want to improve clip 2, passing the request id of clip 3 as a next_request_id (and that of clip 1 as a previous_request_id) will help maintain natural flow in the combined speech. The results will be best when the same model is used across the generations. In case both next_text and next_request_ids is send, next_text will be ignored. A maximum of 3 request_ids can be send. use_pvc_as_ivc : typing.Optional[bool] If true, we won't use PVC version of the voice for the generation but the IVC version. This is a temporary workaround for higher latency in PVC versions. apply_text_normalization : typing.Optional[BodyTextToSpeechFullApplyTextNormalization] This parameter controls text normalization with three modes: 'auto', 'on', and 'off'. When set to 'auto', the system will automatically decide whether to apply text normalization (e.g., spelling out numbers). With 'on', text normalization will always be applied, while with 'off', it will be skipped. apply_language_text_normalization : typing.Optional[bool] This parameter controls language text normalization. This helps with proper pronunciation of text in some supported languages. WARNING: This parameter can heavily increase the latency of the request. Currently only supported for Japanese. request_options : typing.Optional[RequestOptions] Request-specific configuration. You can pass in configuration such as `chunk_size`, and more to customize the request and response. Returns ------- typing.Iterator[HttpResponse[typing.Iterator[bytes]]] The generated audio file """ with self._client_wrapper.httpx_client.stream( f"v1/text-to-speech/{jsonable_encoder(voice_id)}", method="POST", params={ "enable_logging": enable_logging, "optimize_streaming_latency": optimize_streaming_latency, "output_format": output_format, }, json={ "text": text, "model_id": model_id, "language_code": language_code, "voice_settings": convert_and_respect_annotation_metadata( object_=voice_settings, annotation=VoiceSettings, direction="write" ), "pronunciation_dictionary_locators": convert_and_respect_annotation_metadata( object_=pronunciation_dictionary_locators, annotation=typing.Sequence[PronunciationDictionaryVersionLocator], direction="write", ), "seed": seed, "previous_text": previous_text, "next_text": next_text, "previous_request_ids": previous_request_ids, "next_request_ids": next_request_ids, "use_pvc_as_ivc": use_pvc_as_ivc, "apply_text_normalization": apply_text_normalization, "apply_language_text_normalization": apply_language_text_normalization, }, headers={ "content-type": "application/json", }, request_options=request_options, omit=OMIT, ) as _response: def _stream() -> HttpResponse[typing.Iterator[bytes]]: try: if 200 <= _response.status_code < 300: _chunk_size = request_options.get("chunk_size", 1024) if request_options is not None else 1024 return HttpResponse( response=_response, data=(_chunk for _chunk in _response.iter_bytes(chunk_size=_chunk_size)) ) _response.read() if _response.status_code == 422: raise UnprocessableEntityError( headers=dict(_response.headers), body=typing.cast( HttpValidationError, construct_type( type_=HttpValidationError, # type: ignore object_=_response.json(), ), ), ) _response_json = _response.json() except JSONDecodeError: raise ApiError( status_code=_response.status_code, headers=dict(_response.headers), body=_response.text ) raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response_json) yield _stream() def convert_with_timestamps( self, voice_id: str, *, text: str, enable_logging: typing.Optional[bool] = None, optimize_streaming_latency: typing.Optional[int] = None, output_format: typing.Optional[TextToSpeechConvertWithTimestampsRequestOutputFormat] = None, model_id: typing.Optional[str] = OMIT, language_code: typing.Optional[str] = OMIT, voice_settings: typing.Optional[VoiceSettings] = OMIT, pronunciation_dictionary_locators: typing.Optional[ typing.Sequence[PronunciationDictionaryVersionLocator] ] = OMIT, seed: typing.Optional[int] = OMIT, previous_text: typing.Optional[str] = OMIT, next_text: typing.Optional[str] = OMIT, previous_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, next_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, use_pvc_as_ivc: typing.Optional[bool] = OMIT, apply_text_normalization: typing.Optional[BodyTextToSpeechFullWithTimestampsApplyTextNormalization] = OMIT, apply_language_text_normalization: typing.Optional[bool] = OMIT, request_options: typing.Optional[RequestOptions] = None, ) -> HttpResponse[AudioWithTimestampsResponse]: """ Generate speech from text with precise character-level timing information for audio-text synchronization. Parameters ---------- voice_id : str Voice ID to be used, you can use https://api.elevenlabs.io/v1/voices to list all the available voices. text : str The text that will get converted into speech. enable_logging : typing.Optional[bool] When enable_logging is set to false zero retention mode will be used for the request. This will mean history features are unavailable for this request, including request stitching. Zero retention mode may only be used by enterprise customers. optimize_streaming_latency : typing.Optional[int] You can turn on latency optimizations at some cost of quality. The best possible final latency varies by model. Possible values: 0 - default mode (no latency optimizations) 1 - normal latency optimizations (about 50% of possible latency improvement of option 3) 2 - strong latency optimizations (about 75% of possible latency improvement of option 3) 3 - max latency optimizations 4 - max latency optimizations, but also with text normalizer turned off for even more latency savings (best latency, but can mispronounce eg numbers and dates). Defaults to None. output_format : typing.Optional[TextToSpeechConvertWithTimestampsRequestOutputFormat] Output format of the generated audio. Formatted as codec_sample_rate_bitrate. So an mp3 with 22.05kHz sample rate at 32kbs is represented as mp3_22050_32. MP3 with 192kbps bitrate requires you to be subscribed to Creator tier or above. PCM with 44.1kHz sample rate requires you to be subscribed to Pro tier or above. Note that the μ-law format (sometimes written mu-law, often approximated as u-law) is commonly used for Twilio audio inputs. model_id : typing.Optional[str] Identifier of the model that will be used, you can query them using GET /v1/models. The model needs to have support for text to speech, you can check this using the can_do_text_to_speech property. language_code : typing.Optional[str] Language code (ISO 639-1) used to enforce a language for the model and text normalization. If the model does not support provided language code, an error will be returned. voice_settings : typing.Optional[VoiceSettings] Voice settings overriding stored settings for the given voice. They are applied only on the given request. pronunciation_dictionary_locators : typing.Optional[typing.Sequence[PronunciationDictionaryVersionLocator]] A list of pronunciation dictionary locators (id, version_id) to be applied to the text. They will be applied in order. You may have up to 3 locators per request seed : typing.Optional[int] If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed. Must be integer between 0 and 4294967295. previous_text : typing.Optional[str] The text that came before the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. next_text : typing.Optional[str] The text that comes after the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. previous_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that were generated before this generation. Can be used to improve the speech's continuity when splitting up a large task into multiple requests. The results will be best when the same model is used across the generations. In case both previous_text and previous_request_ids is send, previous_text will be ignored. A maximum of 3 request_ids can be send. next_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that come after this generation. next_request_ids is especially useful for maintaining the speech's continuity when regenerating a sample that has had some audio quality issues. For example, if you have generated 3 speech clips, and you want to improve clip 2, passing the request id of clip 3 as a next_request_id (and that of clip 1 as a previous_request_id) will help maintain natural flow in the combined speech. The results will be best when the same model is used across the generations. In case both next_text and next_request_ids is send, next_text will be ignored. A maximum of 3 request_ids can be send. use_pvc_as_ivc : typing.Optional[bool] If true, we won't use PVC version of the voice for the generation but the IVC version. This is a temporary workaround for higher latency in PVC versions. apply_text_normalization : typing.Optional[BodyTextToSpeechFullWithTimestampsApplyTextNormalization] This parameter controls text normalization with three modes: 'auto', 'on', and 'off'. When set to 'auto', the system will automatically decide whether to apply text normalization (e.g., spelling out numbers). With 'on', text normalization will always be applied, while with 'off', it will be skipped. apply_language_text_normalization : typing.Optional[bool] This parameter controls language text normalization. This helps with proper pronunciation of text in some supported languages. WARNING: This parameter can heavily increase the latency of the request. Currently only supported for Japanese. request_options : typing.Optional[RequestOptions] Request-specific configuration. Returns ------- HttpResponse[AudioWithTimestampsResponse] Successful Response """ _response = self._client_wrapper.httpx_client.request( f"v1/text-to-speech/{jsonable_encoder(voice_id)}/with-timestamps", method="POST", params={ "enable_logging": enable_logging, "optimize_streaming_latency": optimize_streaming_latency, "output_format": output_format, }, json={ "text": text, "model_id": model_id, "language_code": language_code, "voice_settings": convert_and_respect_annotation_metadata( object_=voice_settings, annotation=VoiceSettings, direction="write" ), "pronunciation_dictionary_locators": convert_and_respect_annotation_metadata( object_=pronunciation_dictionary_locators, annotation=typing.Sequence[PronunciationDictionaryVersionLocator], direction="write", ), "seed": seed, "previous_text": previous_text, "next_text": next_text, "previous_request_ids": previous_request_ids, "next_request_ids": next_request_ids, "use_pvc_as_ivc": use_pvc_as_ivc, "apply_text_normalization": apply_text_normalization, "apply_language_text_normalization": apply_language_text_normalization, }, headers={ "content-type": "application/json", }, request_options=request_options, omit=OMIT, ) try: if 200 <= _response.status_code < 300: _data = typing.cast( AudioWithTimestampsResponse, construct_type( type_=AudioWithTimestampsResponse, # type: ignore object_=_response.json(), ), ) return HttpResponse(response=_response, data=_data) if _response.status_code == 422: raise UnprocessableEntityError( headers=dict(_response.headers), body=typing.cast( HttpValidationError, construct_type( type_=HttpValidationError, # type: ignore object_=_response.json(), ), ), ) _response_json = _response.json() except JSONDecodeError: raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response.text) raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response_json) @contextlib.contextmanager def stream( self, voice_id: str, *, text: str, enable_logging: typing.Optional[bool] = None, optimize_streaming_latency: typing.Optional[int] = None, output_format: typing.Optional[TextToSpeechStreamRequestOutputFormat] = None, model_id: typing.Optional[str] = OMIT, language_code: typing.Optional[str] = OMIT, voice_settings: typing.Optional[VoiceSettings] = OMIT, pronunciation_dictionary_locators: typing.Optional[ typing.Sequence[PronunciationDictionaryVersionLocator] ] = OMIT, seed: typing.Optional[int] = OMIT, previous_text: typing.Optional[str] = OMIT, next_text: typing.Optional[str] = OMIT, previous_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, next_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, use_pvc_as_ivc: typing.Optional[bool] = OMIT, apply_text_normalization: typing.Optional[BodyTextToSpeechStreamApplyTextNormalization] = OMIT, apply_language_text_normalization: typing.Optional[bool] = OMIT, request_options: typing.Optional[RequestOptions] = None, ) -> typing.Iterator[HttpResponse[typing.Iterator[bytes]]]: """ Converts text into speech using a voice of your choice and returns audio as an audio stream. Parameters ---------- voice_id : str ID of the voice to be used. Use the [Get voices](/docs/api-reference/voices/search) endpoint list all the available voices. text : str The text that will get converted into speech. enable_logging : typing.Optional[bool] When enable_logging is set to false zero retention mode will be used for the request. This will mean history features are unavailable for this request, including request stitching. Zero retention mode may only be used by enterprise customers. optimize_streaming_latency : typing.Optional[int] You can turn on latency optimizations at some cost of quality. The best possible final latency varies by model. Possible values: 0 - default mode (no latency optimizations) 1 - normal latency optimizations (about 50% of possible latency improvement of option 3) 2 - strong latency optimizations (about 75% of possible latency improvement of option 3) 3 - max latency optimizations 4 - max latency optimizations, but also with text normalizer turned off for even more latency savings (best latency, but can mispronounce eg numbers and dates). Defaults to None. output_format : typing.Optional[TextToSpeechStreamRequestOutputFormat] Output format of the generated audio. Formatted as codec_sample_rate_bitrate. So an mp3 with 22.05kHz sample rate at 32kbs is represented as mp3_22050_32. MP3 with 192kbps bitrate requires you to be subscribed to Creator tier or above. PCM with 44.1kHz sample rate requires you to be subscribed to Pro tier or above. Note that the μ-law format (sometimes written mu-law, often approximated as u-law) is commonly used for Twilio audio inputs. model_id : typing.Optional[str] Identifier of the model that will be used, you can query them using GET /v1/models. The model needs to have support for text to speech, you can check this using the can_do_text_to_speech property. language_code : typing.Optional[str] Language code (ISO 639-1) used to enforce a language for the model and text normalization. If the model does not support provided language code, an error will be returned. voice_settings : typing.Optional[VoiceSettings] Voice settings overriding stored settings for the given voice. They are applied only on the given request. pronunciation_dictionary_locators : typing.Optional[typing.Sequence[PronunciationDictionaryVersionLocator]] A list of pronunciation dictionary locators (id, version_id) to be applied to the text. They will be applied in order. You may have up to 3 locators per request seed : typing.Optional[int] If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed. Must be integer between 0 and 4294967295. previous_text : typing.Optional[str] The text that came before the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. next_text : typing.Optional[str] The text that comes after the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. previous_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that were generated before this generation. Can be used to improve the speech's continuity when splitting up a large task into multiple requests. The results will be best when the same model is used across the generations. In case both previous_text and previous_request_ids is send, previous_text will be ignored. A maximum of 3 request_ids can be send. next_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that come after this generation. next_request_ids is especially useful for maintaining the speech's continuity when regenerating a sample that has had some audio quality issues. For example, if you have generated 3 speech clips, and you want to improve clip 2, passing the request id of clip 3 as a next_request_id (and that of clip 1 as a previous_request_id) will help maintain natural flow in the combined speech. The results will be best when the same model is used across the generations. In case both next_text and next_request_ids is send, next_text will be ignored. A maximum of 3 request_ids can be send. use_pvc_as_ivc : typing.Optional[bool] If true, we won't use PVC version of the voice for the generation but the IVC version. This is a temporary workaround for higher latency in PVC versions. apply_text_normalization : typing.Optional[BodyTextToSpeechStreamApplyTextNormalization] This parameter controls text normalization with three modes: 'auto', 'on', and 'off'. When set to 'auto', the system will automatically decide whether to apply text normalization (e.g., spelling out numbers). With 'on', text normalization will always be applied, while with 'off', it will be skipped. apply_language_text_normalization : typing.Optional[bool] This parameter controls language text normalization. This helps with proper pronunciation of text in some supported languages. WARNING: This parameter can heavily increase the latency of the request. Currently only supported for Japanese. request_options : typing.Optional[RequestOptions] Request-specific configuration. You can pass in configuration such as `chunk_size`, and more to customize the request and response. Returns ------- typing.Iterator[HttpResponse[typing.Iterator[bytes]]] Streaming audio data """ with self._client_wrapper.httpx_client.stream( f"v1/text-to-speech/{jsonable_encoder(voice_id)}/stream", method="POST", params={ "enable_logging": enable_logging, "optimize_streaming_latency": optimize_streaming_latency, "output_format": output_format, }, json={ "text": text, "model_id": model_id, "language_code": language_code, "voice_settings": convert_and_respect_annotation_metadata( object_=voice_settings, annotation=VoiceSettings, direction="write" ), "pronunciation_dictionary_locators": convert_and_respect_annotation_metadata( object_=pronunciation_dictionary_locators, annotation=typing.Sequence[PronunciationDictionaryVersionLocator], direction="write", ), "seed": seed, "previous_text": previous_text, "next_text": next_text, "previous_request_ids": previous_request_ids, "next_request_ids": next_request_ids, "use_pvc_as_ivc": use_pvc_as_ivc, "apply_text_normalization": apply_text_normalization, "apply_language_text_normalization": apply_language_text_normalization, }, headers={ "content-type": "application/json", }, request_options=request_options, omit=OMIT, ) as _response: def _stream() -> HttpResponse[typing.Iterator[bytes]]: try: if 200 <= _response.status_code < 300: _chunk_size = request_options.get("chunk_size", 1024) if request_options is not None else 1024 return HttpResponse( response=_response, data=(_chunk for _chunk in _response.iter_bytes(chunk_size=_chunk_size)) ) _response.read() if _response.status_code == 422: raise UnprocessableEntityError( headers=dict(_response.headers), body=typing.cast( HttpValidationError, construct_type( type_=HttpValidationError, # type: ignore object_=_response.json(), ), ), ) _response_json = _response.json() except JSONDecodeError: raise ApiError( status_code=_response.status_code, headers=dict(_response.headers), body=_response.text ) raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response_json) yield _stream() @contextlib.contextmanager def stream_with_timestamps( self, voice_id: str, *, text: str, enable_logging: typing.Optional[bool] = None, optimize_streaming_latency: typing.Optional[int] = None, output_format: typing.Optional[TextToSpeechStreamWithTimestampsRequestOutputFormat] = None, model_id: typing.Optional[str] = OMIT, language_code: typing.Optional[str] = OMIT, voice_settings: typing.Optional[VoiceSettings] = OMIT, pronunciation_dictionary_locators: typing.Optional[ typing.Sequence[PronunciationDictionaryVersionLocator] ] = OMIT, seed: typing.Optional[int] = OMIT, previous_text: typing.Optional[str] = OMIT, next_text: typing.Optional[str] = OMIT, previous_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, next_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, use_pvc_as_ivc: typing.Optional[bool] = OMIT, apply_text_normalization: typing.Optional[BodyTextToSpeechStreamWithTimestampsApplyTextNormalization] = OMIT, apply_language_text_normalization: typing.Optional[bool] = OMIT, request_options: typing.Optional[RequestOptions] = None, ) -> typing.Iterator[HttpResponse[typing.Iterator[StreamingAudioChunkWithTimestampsResponse]]]: """ Converts text into speech using a voice of your choice and returns a stream of JSONs containing audio as a base64 encoded string together with information on when which character was spoken. Parameters ---------- voice_id : str ID of the voice to be used. Use the [Get voices](/docs/api-reference/voices/search) endpoint list all the available voices. text : str The text that will get converted into speech. enable_logging : typing.Optional[bool] When enable_logging is set to false zero retention mode will be used for the request. This will mean history features are unavailable for this request, including request stitching. Zero retention mode may only be used by enterprise customers. optimize_streaming_latency : typing.Optional[int] You can turn on latency optimizations at some cost of quality. The best possible final latency varies by model. Possible values: 0 - default mode (no latency optimizations) 1 - normal latency optimizations (about 50% of possible latency improvement of option 3) 2 - strong latency optimizations (about 75% of possible latency improvement of option 3) 3 - max latency optimizations 4 - max latency optimizations, but also with text normalizer turned off for even more latency savings (best latency, but can mispronounce eg numbers and dates). Defaults to None. output_format : typing.Optional[TextToSpeechStreamWithTimestampsRequestOutputFormat] Output format of the generated audio. Formatted as codec_sample_rate_bitrate. So an mp3 with 22.05kHz sample rate at 32kbs is represented as mp3_22050_32. MP3 with 192kbps bitrate requires you to be subscribed to Creator tier or above. PCM with 44.1kHz sample rate requires you to be subscribed to Pro tier or above. Note that the μ-law format (sometimes written mu-law, often approximated as u-law) is commonly used for Twilio audio inputs. model_id : typing.Optional[str] Identifier of the model that will be used, you can query them using GET /v1/models. The model needs to have support for text to speech, you can check this using the can_do_text_to_speech property. language_code : typing.Optional[str] Language code (ISO 639-1) used to enforce a language for the model and text normalization. If the model does not support provided language code, an error will be returned. voice_settings : typing.Optional[VoiceSettings] Voice settings overriding stored settings for the given voice. They are applied only on the given request. pronunciation_dictionary_locators : typing.Optional[typing.Sequence[PronunciationDictionaryVersionLocator]] A list of pronunciation dictionary locators (id, version_id) to be applied to the text. They will be applied in order. You may have up to 3 locators per request seed : typing.Optional[int] If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed. Must be integer between 0 and 4294967295. previous_text : typing.Optional[str] The text that came before the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. next_text : typing.Optional[str] The text that comes after the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. previous_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that were generated before this generation. Can be used to improve the speech's continuity when splitting up a large task into multiple requests. The results will be best when the same model is used across the generations. In case both previous_text and previous_request_ids is send, previous_text will be ignored. A maximum of 3 request_ids can be send. next_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that come after this generation. next_request_ids is especially useful for maintaining the speech's continuity when regenerating a sample that has had some audio quality issues. For example, if you have generated 3 speech clips, and you want to improve clip 2, passing the request id of clip 3 as a next_request_id (and that of clip 1 as a previous_request_id) will help maintain natural flow in the combined speech. The results will be best when the same model is used across the generations. In case both next_text and next_request_ids is send, next_text will be ignored. A maximum of 3 request_ids can be send. use_pvc_as_ivc : typing.Optional[bool] If true, we won't use PVC version of the voice for the generation but the IVC version. This is a temporary workaround for higher latency in PVC versions. apply_text_normalization : typing.Optional[BodyTextToSpeechStreamWithTimestampsApplyTextNormalization] This parameter controls text normalization with three modes: 'auto', 'on', and 'off'. When set to 'auto', the system will automatically decide whether to apply text normalization (e.g., spelling out numbers). With 'on', text normalization will always be applied, while with 'off', it will be skipped. apply_language_text_normalization : typing.Optional[bool] This parameter controls language text normalization. This helps with proper pronunciation of text in some supported languages. WARNING: This parameter can heavily increase the latency of the request. Currently only supported for Japanese. request_options : typing.Optional[RequestOptions] Request-specific configuration. Yields ------ typing.Iterator[HttpResponse[typing.Iterator[StreamingAudioChunkWithTimestampsResponse]]] Stream of transcription chunks """ with self._client_wrapper.httpx_client.stream( f"v1/text-to-speech/{jsonable_encoder(voice_id)}/stream/with-timestamps", method="POST", params={ "enable_logging": enable_logging, "optimize_streaming_latency": optimize_streaming_latency, "output_format": output_format, }, json={ "text": text, "model_id": model_id, "language_code": language_code, "voice_settings": convert_and_respect_annotation_metadata( object_=voice_settings, annotation=VoiceSettings, direction="write" ), "pronunciation_dictionary_locators": convert_and_respect_annotation_metadata( object_=pronunciation_dictionary_locators, annotation=typing.Sequence[PronunciationDictionaryVersionLocator], direction="write", ), "seed": seed, "previous_text": previous_text, "next_text": next_text, "previous_request_ids": previous_request_ids, "next_request_ids": next_request_ids, "use_pvc_as_ivc": use_pvc_as_ivc, "apply_text_normalization": apply_text_normalization, "apply_language_text_normalization": apply_language_text_normalization, }, headers={ "content-type": "application/json", }, request_options=request_options, omit=OMIT, ) as _response: def _stream() -> HttpResponse[typing.Iterator[StreamingAudioChunkWithTimestampsResponse]]: try: if 200 <= _response.status_code < 300: def _iter(): for _text in _response.iter_lines(): try: if len(_text) == 0: continue yield typing.cast( StreamingAudioChunkWithTimestampsResponse, construct_type( type_=StreamingAudioChunkWithTimestampsResponse, # type: ignore object_=json.loads(_text), ), ) except Exception: pass return return HttpResponse(response=_response, data=_iter()) _response.read() if _response.status_code == 422: raise UnprocessableEntityError( headers=dict(_response.headers), body=typing.cast( HttpValidationError, construct_type( type_=HttpValidationError, # type: ignore object_=_response.json(), ), ), ) _response_json = _response.json() except JSONDecodeError: raise ApiError( status_code=_response.status_code, headers=dict(_response.headers), body=_response.text ) raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response_json) yield _stream() class AsyncRawTextToSpeechClient: def __init__(self, *, client_wrapper: AsyncClientWrapper): self._client_wrapper = client_wrapper @contextlib.asynccontextmanager async def convert( self, voice_id: str, *, text: str, enable_logging: typing.Optional[bool] = None, optimize_streaming_latency: typing.Optional[int] = None, output_format: typing.Optional[TextToSpeechConvertRequestOutputFormat] = None, model_id: typing.Optional[str] = OMIT, language_code: typing.Optional[str] = OMIT, voice_settings: typing.Optional[VoiceSettings] = OMIT, pronunciation_dictionary_locators: typing.Optional[ typing.Sequence[PronunciationDictionaryVersionLocator] ] = OMIT, seed: typing.Optional[int] = OMIT, previous_text: typing.Optional[str] = OMIT, next_text: typing.Optional[str] = OMIT, previous_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, next_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, use_pvc_as_ivc: typing.Optional[bool] = OMIT, apply_text_normalization: typing.Optional[BodyTextToSpeechFullApplyTextNormalization] = OMIT, apply_language_text_normalization: typing.Optional[bool] = OMIT, request_options: typing.Optional[RequestOptions] = None, ) -> typing.AsyncIterator[AsyncHttpResponse[typing.AsyncIterator[bytes]]]: """ Converts text into speech using a voice of your choice and returns audio. Parameters ---------- voice_id : str ID of the voice to be used. Use the [Get voices](/docs/api-reference/voices/search) endpoint list all the available voices. text : str The text that will get converted into speech. enable_logging : typing.Optional[bool] When enable_logging is set to false zero retention mode will be used for the request. This will mean history features are unavailable for this request, including request stitching. Zero retention mode may only be used by enterprise customers. optimize_streaming_latency : typing.Optional[int] You can turn on latency optimizations at some cost of quality. The best possible final latency varies by model. Possible values: 0 - default mode (no latency optimizations) 1 - normal latency optimizations (about 50% of possible latency improvement of option 3) 2 - strong latency optimizations (about 75% of possible latency improvement of option 3) 3 - max latency optimizations 4 - max latency optimizations, but also with text normalizer turned off for even more latency savings (best latency, but can mispronounce eg numbers and dates). Defaults to None. output_format : typing.Optional[TextToSpeechConvertRequestOutputFormat] Output format of the generated audio. Formatted as codec_sample_rate_bitrate. So an mp3 with 22.05kHz sample rate at 32kbs is represented as mp3_22050_32. MP3 with 192kbps bitrate requires you to be subscribed to Creator tier or above. PCM with 44.1kHz sample rate requires you to be subscribed to Pro tier or above. Note that the μ-law format (sometimes written mu-law, often approximated as u-law) is commonly used for Twilio audio inputs. model_id : typing.Optional[str] Identifier of the model that will be used, you can query them using GET /v1/models. The model needs to have support for text to speech, you can check this using the can_do_text_to_speech property. language_code : typing.Optional[str] Language code (ISO 639-1) used to enforce a language for the model and text normalization. If the model does not support provided language code, an error will be returned. voice_settings : typing.Optional[VoiceSettings] Voice settings overriding stored settings for the given voice. They are applied only on the given request. pronunciation_dictionary_locators : typing.Optional[typing.Sequence[PronunciationDictionaryVersionLocator]] A list of pronunciation dictionary locators (id, version_id) to be applied to the text. They will be applied in order. You may have up to 3 locators per request seed : typing.Optional[int] If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed. Must be integer between 0 and 4294967295. previous_text : typing.Optional[str] The text that came before the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. next_text : typing.Optional[str] The text that comes after the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. previous_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that were generated before this generation. Can be used to improve the speech's continuity when splitting up a large task into multiple requests. The results will be best when the same model is used across the generations. In case both previous_text and previous_request_ids is send, previous_text will be ignored. A maximum of 3 request_ids can be send. next_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that come after this generation. next_request_ids is especially useful for maintaining the speech's continuity when regenerating a sample that has had some audio quality issues. For example, if you have generated 3 speech clips, and you want to improve clip 2, passing the request id of clip 3 as a next_request_id (and that of clip 1 as a previous_request_id) will help maintain natural flow in the combined speech. The results will be best when the same model is used across the generations. In case both next_text and next_request_ids is send, next_text will be ignored. A maximum of 3 request_ids can be send. use_pvc_as_ivc : typing.Optional[bool] If true, we won't use PVC version of the voice for the generation but the IVC version. This is a temporary workaround for higher latency in PVC versions. apply_text_normalization : typing.Optional[BodyTextToSpeechFullApplyTextNormalization] This parameter controls text normalization with three modes: 'auto', 'on', and 'off'. When set to 'auto', the system will automatically decide whether to apply text normalization (e.g., spelling out numbers). With 'on', text normalization will always be applied, while with 'off', it will be skipped. apply_language_text_normalization : typing.Optional[bool] This parameter controls language text normalization. This helps with proper pronunciation of text in some supported languages. WARNING: This parameter can heavily increase the latency of the request. Currently only supported for Japanese. request_options : typing.Optional[RequestOptions] Request-specific configuration. You can pass in configuration such as `chunk_size`, and more to customize the request and response. Returns ------- typing.AsyncIterator[AsyncHttpResponse[typing.AsyncIterator[bytes]]] The generated audio file """ async with self._client_wrapper.httpx_client.stream( f"v1/text-to-speech/{jsonable_encoder(voice_id)}", method="POST", params={ "enable_logging": enable_logging, "optimize_streaming_latency": optimize_streaming_latency, "output_format": output_format, }, json={ "text": text, "model_id": model_id, "language_code": language_code, "voice_settings": convert_and_respect_annotation_metadata( object_=voice_settings, annotation=VoiceSettings, direction="write" ), "pronunciation_dictionary_locators": convert_and_respect_annotation_metadata( object_=pronunciation_dictionary_locators, annotation=typing.Sequence[PronunciationDictionaryVersionLocator], direction="write", ), "seed": seed, "previous_text": previous_text, "next_text": next_text, "previous_request_ids": previous_request_ids, "next_request_ids": next_request_ids, "use_pvc_as_ivc": use_pvc_as_ivc, "apply_text_normalization": apply_text_normalization, "apply_language_text_normalization": apply_language_text_normalization, }, headers={ "content-type": "application/json", }, request_options=request_options, omit=OMIT, ) as _response: async def _stream() -> AsyncHttpResponse[typing.AsyncIterator[bytes]]: try: if 200 <= _response.status_code < 300: _chunk_size = request_options.get("chunk_size", 1024) if request_options is not None else 1024 return AsyncHttpResponse( response=_response, data=(_chunk async for _chunk in _response.aiter_bytes(chunk_size=_chunk_size)), ) await _response.aread() if _response.status_code == 422: raise UnprocessableEntityError( headers=dict(_response.headers), body=typing.cast( HttpValidationError, construct_type( type_=HttpValidationError, # type: ignore object_=_response.json(), ), ), ) _response_json = _response.json() except JSONDecodeError: raise ApiError( status_code=_response.status_code, headers=dict(_response.headers), body=_response.text ) raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response_json) yield await _stream() async def convert_with_timestamps( self, voice_id: str, *, text: str, enable_logging: typing.Optional[bool] = None, optimize_streaming_latency: typing.Optional[int] = None, output_format: typing.Optional[TextToSpeechConvertWithTimestampsRequestOutputFormat] = None, model_id: typing.Optional[str] = OMIT, language_code: typing.Optional[str] = OMIT, voice_settings: typing.Optional[VoiceSettings] = OMIT, pronunciation_dictionary_locators: typing.Optional[ typing.Sequence[PronunciationDictionaryVersionLocator] ] = OMIT, seed: typing.Optional[int] = OMIT, previous_text: typing.Optional[str] = OMIT, next_text: typing.Optional[str] = OMIT, previous_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, next_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, use_pvc_as_ivc: typing.Optional[bool] = OMIT, apply_text_normalization: typing.Optional[BodyTextToSpeechFullWithTimestampsApplyTextNormalization] = OMIT, apply_language_text_normalization: typing.Optional[bool] = OMIT, request_options: typing.Optional[RequestOptions] = None, ) -> AsyncHttpResponse[AudioWithTimestampsResponse]: """ Generate speech from text with precise character-level timing information for audio-text synchronization. Parameters ---------- voice_id : str Voice ID to be used, you can use https://api.elevenlabs.io/v1/voices to list all the available voices. text : str The text that will get converted into speech. enable_logging : typing.Optional[bool] When enable_logging is set to false zero retention mode will be used for the request. This will mean history features are unavailable for this request, including request stitching. Zero retention mode may only be used by enterprise customers. optimize_streaming_latency : typing.Optional[int] You can turn on latency optimizations at some cost of quality. The best possible final latency varies by model. Possible values: 0 - default mode (no latency optimizations) 1 - normal latency optimizations (about 50% of possible latency improvement of option 3) 2 - strong latency optimizations (about 75% of possible latency improvement of option 3) 3 - max latency optimizations 4 - max latency optimizations, but also with text normalizer turned off for even more latency savings (best latency, but can mispronounce eg numbers and dates). Defaults to None. output_format : typing.Optional[TextToSpeechConvertWithTimestampsRequestOutputFormat] Output format of the generated audio. Formatted as codec_sample_rate_bitrate. So an mp3 with 22.05kHz sample rate at 32kbs is represented as mp3_22050_32. MP3 with 192kbps bitrate requires you to be subscribed to Creator tier or above. PCM with 44.1kHz sample rate requires you to be subscribed to Pro tier or above. Note that the μ-law format (sometimes written mu-law, often approximated as u-law) is commonly used for Twilio audio inputs. model_id : typing.Optional[str] Identifier of the model that will be used, you can query them using GET /v1/models. The model needs to have support for text to speech, you can check this using the can_do_text_to_speech property. language_code : typing.Optional[str] Language code (ISO 639-1) used to enforce a language for the model and text normalization. If the model does not support provided language code, an error will be returned. voice_settings : typing.Optional[VoiceSettings] Voice settings overriding stored settings for the given voice. They are applied only on the given request. pronunciation_dictionary_locators : typing.Optional[typing.Sequence[PronunciationDictionaryVersionLocator]] A list of pronunciation dictionary locators (id, version_id) to be applied to the text. They will be applied in order. You may have up to 3 locators per request seed : typing.Optional[int] If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed. Must be integer between 0 and 4294967295. previous_text : typing.Optional[str] The text that came before the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. next_text : typing.Optional[str] The text that comes after the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. previous_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that were generated before this generation. Can be used to improve the speech's continuity when splitting up a large task into multiple requests. The results will be best when the same model is used across the generations. In case both previous_text and previous_request_ids is send, previous_text will be ignored. A maximum of 3 request_ids can be send. next_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that come after this generation. next_request_ids is especially useful for maintaining the speech's continuity when regenerating a sample that has had some audio quality issues. For example, if you have generated 3 speech clips, and you want to improve clip 2, passing the request id of clip 3 as a next_request_id (and that of clip 1 as a previous_request_id) will help maintain natural flow in the combined speech. The results will be best when the same model is used across the generations. In case both next_text and next_request_ids is send, next_text will be ignored. A maximum of 3 request_ids can be send. use_pvc_as_ivc : typing.Optional[bool] If true, we won't use PVC version of the voice for the generation but the IVC version. This is a temporary workaround for higher latency in PVC versions. apply_text_normalization : typing.Optional[BodyTextToSpeechFullWithTimestampsApplyTextNormalization] This parameter controls text normalization with three modes: 'auto', 'on', and 'off'. When set to 'auto', the system will automatically decide whether to apply text normalization (e.g., spelling out numbers). With 'on', text normalization will always be applied, while with 'off', it will be skipped. apply_language_text_normalization : typing.Optional[bool] This parameter controls language text normalization. This helps with proper pronunciation of text in some supported languages. WARNING: This parameter can heavily increase the latency of the request. Currently only supported for Japanese. request_options : typing.Optional[RequestOptions] Request-specific configuration. Returns ------- AsyncHttpResponse[AudioWithTimestampsResponse] Successful Response """ _response = await self._client_wrapper.httpx_client.request( f"v1/text-to-speech/{jsonable_encoder(voice_id)}/with-timestamps", method="POST", params={ "enable_logging": enable_logging, "optimize_streaming_latency": optimize_streaming_latency, "output_format": output_format, }, json={ "text": text, "model_id": model_id, "language_code": language_code, "voice_settings": convert_and_respect_annotation_metadata( object_=voice_settings, annotation=VoiceSettings, direction="write" ), "pronunciation_dictionary_locators": convert_and_respect_annotation_metadata( object_=pronunciation_dictionary_locators, annotation=typing.Sequence[PronunciationDictionaryVersionLocator], direction="write", ), "seed": seed, "previous_text": previous_text, "next_text": next_text, "previous_request_ids": previous_request_ids, "next_request_ids": next_request_ids, "use_pvc_as_ivc": use_pvc_as_ivc, "apply_text_normalization": apply_text_normalization, "apply_language_text_normalization": apply_language_text_normalization, }, headers={ "content-type": "application/json", }, request_options=request_options, omit=OMIT, ) try: if 200 <= _response.status_code < 300: _data = typing.cast( AudioWithTimestampsResponse, construct_type( type_=AudioWithTimestampsResponse, # type: ignore object_=_response.json(), ), ) return AsyncHttpResponse(response=_response, data=_data) if _response.status_code == 422: raise UnprocessableEntityError( headers=dict(_response.headers), body=typing.cast( HttpValidationError, construct_type( type_=HttpValidationError, # type: ignore object_=_response.json(), ), ), ) _response_json = _response.json() except JSONDecodeError: raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response.text) raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response_json) @contextlib.asynccontextmanager async def stream( self, voice_id: str, *, text: str, enable_logging: typing.Optional[bool] = None, optimize_streaming_latency: typing.Optional[int] = None, output_format: typing.Optional[TextToSpeechStreamRequestOutputFormat] = None, model_id: typing.Optional[str] = OMIT, language_code: typing.Optional[str] = OMIT, voice_settings: typing.Optional[VoiceSettings] = OMIT, pronunciation_dictionary_locators: typing.Optional[ typing.Sequence[PronunciationDictionaryVersionLocator] ] = OMIT, seed: typing.Optional[int] = OMIT, previous_text: typing.Optional[str] = OMIT, next_text: typing.Optional[str] = OMIT, previous_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, next_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, use_pvc_as_ivc: typing.Optional[bool] = OMIT, apply_text_normalization: typing.Optional[BodyTextToSpeechStreamApplyTextNormalization] = OMIT, apply_language_text_normalization: typing.Optional[bool] = OMIT, request_options: typing.Optional[RequestOptions] = None, ) -> typing.AsyncIterator[AsyncHttpResponse[typing.AsyncIterator[bytes]]]: """ Converts text into speech using a voice of your choice and returns audio as an audio stream. Parameters ---------- voice_id : str ID of the voice to be used. Use the [Get voices](/docs/api-reference/voices/search) endpoint list all the available voices. text : str The text that will get converted into speech. enable_logging : typing.Optional[bool] When enable_logging is set to false zero retention mode will be used for the request. This will mean history features are unavailable for this request, including request stitching. Zero retention mode may only be used by enterprise customers. optimize_streaming_latency : typing.Optional[int] You can turn on latency optimizations at some cost of quality. The best possible final latency varies by model. Possible values: 0 - default mode (no latency optimizations) 1 - normal latency optimizations (about 50% of possible latency improvement of option 3) 2 - strong latency optimizations (about 75% of possible latency improvement of option 3) 3 - max latency optimizations 4 - max latency optimizations, but also with text normalizer turned off for even more latency savings (best latency, but can mispronounce eg numbers and dates). Defaults to None. output_format : typing.Optional[TextToSpeechStreamRequestOutputFormat] Output format of the generated audio. Formatted as codec_sample_rate_bitrate. So an mp3 with 22.05kHz sample rate at 32kbs is represented as mp3_22050_32. MP3 with 192kbps bitrate requires you to be subscribed to Creator tier or above. PCM with 44.1kHz sample rate requires you to be subscribed to Pro tier or above. Note that the μ-law format (sometimes written mu-law, often approximated as u-law) is commonly used for Twilio audio inputs. model_id : typing.Optional[str] Identifier of the model that will be used, you can query them using GET /v1/models. The model needs to have support for text to speech, you can check this using the can_do_text_to_speech property. language_code : typing.Optional[str] Language code (ISO 639-1) used to enforce a language for the model and text normalization. If the model does not support provided language code, an error will be returned. voice_settings : typing.Optional[VoiceSettings] Voice settings overriding stored settings for the given voice. They are applied only on the given request. pronunciation_dictionary_locators : typing.Optional[typing.Sequence[PronunciationDictionaryVersionLocator]] A list of pronunciation dictionary locators (id, version_id) to be applied to the text. They will be applied in order. You may have up to 3 locators per request seed : typing.Optional[int] If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed. Must be integer between 0 and 4294967295. previous_text : typing.Optional[str] The text that came before the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. next_text : typing.Optional[str] The text that comes after the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. previous_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that were generated before this generation. Can be used to improve the speech's continuity when splitting up a large task into multiple requests. The results will be best when the same model is used across the generations. In case both previous_text and previous_request_ids is send, previous_text will be ignored. A maximum of 3 request_ids can be send. next_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that come after this generation. next_request_ids is especially useful for maintaining the speech's continuity when regenerating a sample that has had some audio quality issues. For example, if you have generated 3 speech clips, and you want to improve clip 2, passing the request id of clip 3 as a next_request_id (and that of clip 1 as a previous_request_id) will help maintain natural flow in the combined speech. The results will be best when the same model is used across the generations. In case both next_text and next_request_ids is send, next_text will be ignored. A maximum of 3 request_ids can be send. use_pvc_as_ivc : typing.Optional[bool] If true, we won't use PVC version of the voice for the generation but the IVC version. This is a temporary workaround for higher latency in PVC versions. apply_text_normalization : typing.Optional[BodyTextToSpeechStreamApplyTextNormalization] This parameter controls text normalization with three modes: 'auto', 'on', and 'off'. When set to 'auto', the system will automatically decide whether to apply text normalization (e.g., spelling out numbers). With 'on', text normalization will always be applied, while with 'off', it will be skipped. apply_language_text_normalization : typing.Optional[bool] This parameter controls language text normalization. This helps with proper pronunciation of text in some supported languages. WARNING: This parameter can heavily increase the latency of the request. Currently only supported for Japanese. request_options : typing.Optional[RequestOptions] Request-specific configuration. You can pass in configuration such as `chunk_size`, and more to customize the request and response. Returns ------- typing.AsyncIterator[AsyncHttpResponse[typing.AsyncIterator[bytes]]] Streaming audio data """ async with self._client_wrapper.httpx_client.stream( f"v1/text-to-speech/{jsonable_encoder(voice_id)}/stream", method="POST", params={ "enable_logging": enable_logging, "optimize_streaming_latency": optimize_streaming_latency, "output_format": output_format, }, json={ "text": text, "model_id": model_id, "language_code": language_code, "voice_settings": convert_and_respect_annotation_metadata( object_=voice_settings, annotation=VoiceSettings, direction="write" ), "pronunciation_dictionary_locators": convert_and_respect_annotation_metadata( object_=pronunciation_dictionary_locators, annotation=typing.Sequence[PronunciationDictionaryVersionLocator], direction="write", ), "seed": seed, "previous_text": previous_text, "next_text": next_text, "previous_request_ids": previous_request_ids, "next_request_ids": next_request_ids, "use_pvc_as_ivc": use_pvc_as_ivc, "apply_text_normalization": apply_text_normalization, "apply_language_text_normalization": apply_language_text_normalization, }, headers={ "content-type": "application/json", }, request_options=request_options, omit=OMIT, ) as _response: async def _stream() -> AsyncHttpResponse[typing.AsyncIterator[bytes]]: try: if 200 <= _response.status_code < 300: _chunk_size = request_options.get("chunk_size", 1024) if request_options is not None else 1024 return AsyncHttpResponse( response=_response, data=(_chunk async for _chunk in _response.aiter_bytes(chunk_size=_chunk_size)), ) await _response.aread() if _response.status_code == 422: raise UnprocessableEntityError( headers=dict(_response.headers), body=typing.cast( HttpValidationError, construct_type( type_=HttpValidationError, # type: ignore object_=_response.json(), ), ), ) _response_json = _response.json() except JSONDecodeError: raise ApiError( status_code=_response.status_code, headers=dict(_response.headers), body=_response.text ) raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response_json) yield await _stream() @contextlib.asynccontextmanager async def stream_with_timestamps( self, voice_id: str, *, text: str, enable_logging: typing.Optional[bool] = None, optimize_streaming_latency: typing.Optional[int] = None, output_format: typing.Optional[TextToSpeechStreamWithTimestampsRequestOutputFormat] = None, model_id: typing.Optional[str] = OMIT, language_code: typing.Optional[str] = OMIT, voice_settings: typing.Optional[VoiceSettings] = OMIT, pronunciation_dictionary_locators: typing.Optional[ typing.Sequence[PronunciationDictionaryVersionLocator] ] = OMIT, seed: typing.Optional[int] = OMIT, previous_text: typing.Optional[str] = OMIT, next_text: typing.Optional[str] = OMIT, previous_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, next_request_ids: typing.Optional[typing.Sequence[str]] = OMIT, use_pvc_as_ivc: typing.Optional[bool] = OMIT, apply_text_normalization: typing.Optional[BodyTextToSpeechStreamWithTimestampsApplyTextNormalization] = OMIT, apply_language_text_normalization: typing.Optional[bool] = OMIT, request_options: typing.Optional[RequestOptions] = None, ) -> typing.AsyncIterator[AsyncHttpResponse[typing.AsyncIterator[StreamingAudioChunkWithTimestampsResponse]]]: """ Converts text into speech using a voice of your choice and returns a stream of JSONs containing audio as a base64 encoded string together with information on when which character was spoken. Parameters ---------- voice_id : str ID of the voice to be used. Use the [Get voices](/docs/api-reference/voices/search) endpoint list all the available voices. text : str The text that will get converted into speech. enable_logging : typing.Optional[bool] When enable_logging is set to false zero retention mode will be used for the request. This will mean history features are unavailable for this request, including request stitching. Zero retention mode may only be used by enterprise customers. optimize_streaming_latency : typing.Optional[int] You can turn on latency optimizations at some cost of quality. The best possible final latency varies by model. Possible values: 0 - default mode (no latency optimizations) 1 - normal latency optimizations (about 50% of possible latency improvement of option 3) 2 - strong latency optimizations (about 75% of possible latency improvement of option 3) 3 - max latency optimizations 4 - max latency optimizations, but also with text normalizer turned off for even more latency savings (best latency, but can mispronounce eg numbers and dates). Defaults to None. output_format : typing.Optional[TextToSpeechStreamWithTimestampsRequestOutputFormat] Output format of the generated audio. Formatted as codec_sample_rate_bitrate. So an mp3 with 22.05kHz sample rate at 32kbs is represented as mp3_22050_32. MP3 with 192kbps bitrate requires you to be subscribed to Creator tier or above. PCM with 44.1kHz sample rate requires you to be subscribed to Pro tier or above. Note that the μ-law format (sometimes written mu-law, often approximated as u-law) is commonly used for Twilio audio inputs. model_id : typing.Optional[str] Identifier of the model that will be used, you can query them using GET /v1/models. The model needs to have support for text to speech, you can check this using the can_do_text_to_speech property. language_code : typing.Optional[str] Language code (ISO 639-1) used to enforce a language for the model and text normalization. If the model does not support provided language code, an error will be returned. voice_settings : typing.Optional[VoiceSettings] Voice settings overriding stored settings for the given voice. They are applied only on the given request. pronunciation_dictionary_locators : typing.Optional[typing.Sequence[PronunciationDictionaryVersionLocator]] A list of pronunciation dictionary locators (id, version_id) to be applied to the text. They will be applied in order. You may have up to 3 locators per request seed : typing.Optional[int] If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed. Must be integer between 0 and 4294967295. previous_text : typing.Optional[str] The text that came before the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. next_text : typing.Optional[str] The text that comes after the text of the current request. Can be used to improve the speech's continuity when concatenating together multiple generations or to influence the speech's continuity in the current generation. previous_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that were generated before this generation. Can be used to improve the speech's continuity when splitting up a large task into multiple requests. The results will be best when the same model is used across the generations. In case both previous_text and previous_request_ids is send, previous_text will be ignored. A maximum of 3 request_ids can be send. next_request_ids : typing.Optional[typing.Sequence[str]] A list of request_id of the samples that come after this generation. next_request_ids is especially useful for maintaining the speech's continuity when regenerating a sample that has had some audio quality issues. For example, if you have generated 3 speech clips, and you want to improve clip 2, passing the request id of clip 3 as a next_request_id (and that of clip 1 as a previous_request_id) will help maintain natural flow in the combined speech. The results will be best when the same model is used across the generations. In case both next_text and next_request_ids is send, next_text will be ignored. A maximum of 3 request_ids can be send. use_pvc_as_ivc : typing.Optional[bool] If true, we won't use PVC version of the voice for the generation but the IVC version. This is a temporary workaround for higher latency in PVC versions. apply_text_normalization : typing.Optional[BodyTextToSpeechStreamWithTimestampsApplyTextNormalization] This parameter controls text normalization with three modes: 'auto', 'on', and 'off'. When set to 'auto', the system will automatically decide whether to apply text normalization (e.g., spelling out numbers). With 'on', text normalization will always be applied, while with 'off', it will be skipped. apply_language_text_normalization : typing.Optional[bool] This parameter controls language text normalization. This helps with proper pronunciation of text in some supported languages. WARNING: This parameter can heavily increase the latency of the request. Currently only supported for Japanese. request_options : typing.Optional[RequestOptions] Request-specific configuration. Yields ------ typing.AsyncIterator[AsyncHttpResponse[typing.AsyncIterator[StreamingAudioChunkWithTimestampsResponse]]] Stream of transcription chunks """ async with self._client_wrapper.httpx_client.stream( f"v1/text-to-speech/{jsonable_encoder(voice_id)}/stream/with-timestamps", method="POST", params={ "enable_logging": enable_logging, "optimize_streaming_latency": optimize_streaming_latency, "output_format": output_format, }, json={ "text": text, "model_id": model_id, "language_code": language_code, "voice_settings": convert_and_respect_annotation_metadata( object_=voice_settings, annotation=VoiceSettings, direction="write" ), "pronunciation_dictionary_locators": convert_and_respect_annotation_metadata( object_=pronunciation_dictionary_locators, annotation=typing.Sequence[PronunciationDictionaryVersionLocator], direction="write", ), "seed": seed, "previous_text": previous_text, "next_text": next_text, "previous_request_ids": previous_request_ids, "next_request_ids": next_request_ids, "use_pvc_as_ivc": use_pvc_as_ivc, "apply_text_normalization": apply_text_normalization, "apply_language_text_normalization": apply_language_text_normalization, }, headers={ "content-type": "application/json", }, request_options=request_options, omit=OMIT, ) as _response: async def _stream() -> AsyncHttpResponse[typing.AsyncIterator[StreamingAudioChunkWithTimestampsResponse]]: try: if 200 <= _response.status_code < 300: async def _iter(): async for _text in _response.aiter_lines(): try: if len(_text) == 0: continue yield typing.cast( StreamingAudioChunkWithTimestampsResponse, construct_type( type_=StreamingAudioChunkWithTimestampsResponse, # type: ignore object_=json.loads(_text), ), ) except Exception: pass return return AsyncHttpResponse(response=_response, data=_iter()) await _response.aread() if _response.status_code == 422: raise UnprocessableEntityError( headers=dict(_response.headers), body=typing.cast( HttpValidationError, construct_type( type_=HttpValidationError, # type: ignore object_=_response.json(), ), ), ) _response_json = _response.json() except JSONDecodeError: raise ApiError( status_code=_response.status_code, headers=dict(_response.headers), body=_response.text ) raise ApiError(status_code=_response.status_code, headers=dict(_response.headers), body=_response_json) yield await _stream()