openai.resources.audio

The audio module provides classes for audio processing operations like transcription, translation, and speech synthesis.

Use the audio module by accessing the OpenAI.audio attribute on the client object, and then access the attribute for the feature you'd like to use:

OpenAI.audio.transcriptions - Transcribe spoken audio to text
OpenAI.audio.translations - Translate spoken audio to English
OpenAI.audio.speech - Generate spoken audio from text

Examples:

from pathlib import Path

from openai import OpenAI

openai = OpenAI()

speech_file_path = Path(__file__).parent / "speech.mp3"

    # Create text-to-speech audio file
    with openai.audio.speech.with_streaming_response.create(
        model="tts-1",
        voice="alloy",
        input="The quick brown fox jumps over the lazy dog.",
    ) as response:
        response.stream_to_file(speech_file_path)

Modules:

Name	Description
`audio`	The `audio` module provides classes for audio processing operations like transcription, translation, and speech synthesis. Use the `audio` module by accessing the `OpenAI.audio` attribute on the client object, and then access the attribute for the feature you'd like to use: `OpenAI.audio.transcriptions` - Transcribe spoken audio to text `OpenAI.audio.translations` - Translate spoken audio to English `OpenAI.audio.speech` - Generate spoken audio from text Examples: `from pathlib import Path from openai import OpenAI openai = OpenAI() speech_file_path = Path(__file__).parent / "speech.mp3" # Create text-to-speech audio file with openai.audio.speech.with_streaming_response.create( model="tts-1", voice="alloy", input="The quick brown fox jumps over the lazy dog.", ) as response: response.stream_to_file(speech_file_path)`
`speech`
`transcriptions`
`translations`

Classes:

Name	Description
`AsyncAudio`
`AsyncAudioWithRawResponse`
`AsyncAudioWithStreamingResponse`
`AsyncSpeech`
`AsyncSpeechWithRawResponse`
`AsyncSpeechWithStreamingResponse`
`AsyncTranscriptions`
`AsyncTranscriptionsWithRawResponse`
`AsyncTranscriptionsWithStreamingResponse`
`AsyncTranslations`
`AsyncTranslationsWithRawResponse`
`AsyncTranslationsWithStreamingResponse`
`Audio`
`AudioWithRawResponse`
`AudioWithStreamingResponse`
`Speech`
`SpeechWithRawResponse`
`SpeechWithStreamingResponse`
`Transcriptions`
`TranscriptionsWithRawResponse`
`TranscriptionsWithStreamingResponse`
`Translations`
`TranslationsWithRawResponse`
`TranslationsWithStreamingResponse`

AsyncAudio

AsyncAudio(client: AsyncOpenAI)

Methods:

Name	Description
`speech`
`transcriptions`
`translations`
`with_raw_response`
`with_streaming_response`

speech

speech() -> AsyncSpeech

transcriptions

transcriptions() -> AsyncTranscriptions

translations

translations() -> AsyncTranslations

with_raw_response

with_raw_response() -> AsyncAudioWithRawResponse

with_streaming_response

with_streaming_response() -> (
    AsyncAudioWithStreamingResponse
)

AsyncAudioWithRawResponse

AsyncAudioWithRawResponse(audio: AsyncAudio)

Methods:

Name	Description
`speech`
`transcriptions`
`translations`

speech

speech() -> AsyncSpeechWithRawResponse

transcriptions

transcriptions() -> AsyncTranscriptionsWithRawResponse

translations

translations() -> AsyncTranslationsWithRawResponse

AsyncAudioWithStreamingResponse

AsyncAudioWithStreamingResponse(audio: AsyncAudio)

Methods:

Name	Description
`speech`
`transcriptions`
`translations`

speech

speech() -> AsyncSpeechWithStreamingResponse

transcriptions

transcriptions() -> (
    AsyncTranscriptionsWithStreamingResponse
)

translations

translations() -> AsyncTranslationsWithStreamingResponse

AsyncSpeech

AsyncSpeech(client: AsyncOpenAI)

Methods:

Name	Description
`create`	Generates audio from the input text.
`with_raw_response`
`with_streaming_response`

create `async`

create(
    *,
    input: str,
    model: Union[str, Literal["tts-1", "tts-1-hd"]],
    voice: Literal[
        "alloy", "echo", "fable", "onyx", "nova", "shimmer"
    ],
    response_format: (
        Literal["mp3", "opus", "aac", "flac", "wav", "pcm"]
        | NotGiven
    ) = NOT_GIVEN,
    speed: float | NotGiven = NOT_GIVEN,
    extra_headers: Headers | None = None,
    extra_query: Query | None = None,
    extra_body: Body | None = None,
    timeout: float | Timeout | None | NotGiven = NOT_GIVEN
) -> HttpxBinaryResponseContent

Generates audio from the input text.

Parameters:

Name	Type	Description	Default
`input`	`str`	The text to generate audio for. The maximum length is 4096 characters.	required
`model`	`Union[str, Literal['tts-1', 'tts-1-hd']]`	One of the available TTS models: `tts-1` or `tts-1-hd`	required
`voice`	`Literal['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer']`	The voice to use when generating the audio. Supported voices are `alloy`, `echo`, `fable`, `onyx`, `nova`, and `shimmer`. Previews of the voices are available in the Text to speech guide.	required
`response_format`	`Literal['mp3', 'opus', 'aac', 'flac', 'wav', 'pcm'] \| NotGiven`	The format to audio in. Supported formats are `mp3`, `opus`, `aac`, `flac`, `wav`, and `pcm`.	`NOT_GIVEN`
`speed`	`float \| NotGiven`	The speed of the generated audio. Select a value from `0.25` to `4.0`. `1.0` is the default.	`NOT_GIVEN`
`extra_headers`	`Headers \| None`	Send extra headers	`None`
`extra_query`	`Query \| None`	Add additional query parameters to the request	`None`
`extra_body`	`Body \| None`	Add additional JSON properties to the request	`None`
`timeout`	`float \| Timeout \| None \| NotGiven`	Override the client-level default timeout for this request, in seconds	`NOT_GIVEN`

with_raw_response

with_raw_response() -> AsyncSpeechWithRawResponse

with_streaming_response

with_streaming_response() -> (
    AsyncSpeechWithStreamingResponse
)

AsyncSpeechWithRawResponse

AsyncSpeechWithRawResponse(speech: AsyncSpeech)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = async_to_raw_response_wrapper(create)

AsyncSpeechWithStreamingResponse

AsyncSpeechWithStreamingResponse(speech: AsyncSpeech)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = async_to_custom_streamed_response_wrapper(
    create, AsyncStreamedBinaryAPIResponse
)

AsyncTranscriptions

AsyncTranscriptions(client: AsyncOpenAI)

Methods:

Name	Description
`create`	Transcribes audio into the input language.
`with_raw_response`
`with_streaming_response`

create `async`

create(
    *,
    file: FileTypes,
    model: Union[str, Literal["whisper-1"]],
    language: str | NotGiven = NOT_GIVEN,
    prompt: str | NotGiven = NOT_GIVEN,
    response_format: (
        Literal[
            "json", "text", "srt", "verbose_json", "vtt"
        ]
        | NotGiven
    ) = NOT_GIVEN,
    temperature: float | NotGiven = NOT_GIVEN,
    timestamp_granularities: (
        List[Literal["word", "segment"]] | NotGiven
    ) = NOT_GIVEN,
    extra_headers: Headers | None = None,
    extra_query: Query | None = None,
    extra_body: Body | None = None,
    timeout: float | Timeout | None | NotGiven = NOT_GIVEN
) -> Transcription

Transcribes audio into the input language.

Parameters:

Name	Type	Description	Default
`file`	`FileTypes`	The audio file object (not file name) to transcribe, in one of these formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, or webm.	required
`model`	`Union[str, Literal['whisper-1']]`	ID of the model to use. Only `whisper-1` (which is powered by our open source Whisper V2 model) is currently available.	required
`language`	`str \| NotGiven`	The language of the input audio. Supplying the input language in ISO-639-1 format will improve accuracy and latency.	`NOT_GIVEN`
`prompt`	`str \| NotGiven`	An optional text to guide the model's style or continue a previous audio segment. The prompt should match the audio language.	`NOT_GIVEN`
`response_format`	`Literal['json', 'text', 'srt', 'verbose_json', 'vtt'] \| NotGiven`	The format of the transcript output, in one of these options: `json`, `text`, `srt`, `verbose_json`, or `vtt`.	`NOT_GIVEN`
`temperature`	`float \| NotGiven`	The sampling temperature, between 0 and 1. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. If set to 0, the model will use log probability to automatically increase the temperature until certain thresholds are hit.	`NOT_GIVEN`
`timestamp_granularities`	`List[Literal['word', 'segment']] \| NotGiven`	The timestamp granularities to populate for this transcription. `response_format` must be set `verbose_json` to use timestamp granularities. Either or both of these options are supported: `word`, or `segment`. Note: There is no additional latency for segment timestamps, but generating word timestamps incurs additional latency.	`NOT_GIVEN`
`extra_headers`	`Headers \| None`	Send extra headers	`None`
`extra_query`	`Query \| None`	Add additional query parameters to the request	`None`
`extra_body`	`Body \| None`	Add additional JSON properties to the request	`None`
`timeout`	`float \| Timeout \| None \| NotGiven`	Override the client-level default timeout for this request, in seconds	`NOT_GIVEN`

with_raw_response

with_raw_response() -> AsyncTranscriptionsWithRawResponse

with_streaming_response

with_streaming_response() -> (
    AsyncTranscriptionsWithStreamingResponse
)

AsyncTranscriptionsWithRawResponse

AsyncTranscriptionsWithRawResponse(
    transcriptions: AsyncTranscriptions,
)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = async_to_raw_response_wrapper(create)

AsyncTranscriptionsWithStreamingResponse

AsyncTranscriptionsWithStreamingResponse(
    transcriptions: AsyncTranscriptions,
)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = async_to_streamed_response_wrapper(create)

AsyncTranslations

AsyncTranslations(client: AsyncOpenAI)

Methods:

Name	Description
`create`	Translates audio into English.
`with_raw_response`
`with_streaming_response`

create `async`

create(
    *,
    file: FileTypes,
    model: Union[str, Literal["whisper-1"]],
    prompt: str | NotGiven = NOT_GIVEN,
    response_format: str | NotGiven = NOT_GIVEN,
    temperature: float | NotGiven = NOT_GIVEN,
    extra_headers: Headers | None = None,
    extra_query: Query | None = None,
    extra_body: Body | None = None,
    timeout: float | Timeout | None | NotGiven = NOT_GIVEN
) -> Translation

Translates audio into English.

Parameters:

Name	Type	Description	Default
`file`	`FileTypes`	The audio file object (not file name) translate, in one of these formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, or webm.	required
`model`	`Union[str, Literal['whisper-1']]`	ID of the model to use. Only `whisper-1` (which is powered by our open source Whisper V2 model) is currently available.	required
`prompt`	`str \| NotGiven`	An optional text to guide the model's style or continue a previous audio segment. The prompt should be in English.	`NOT_GIVEN`
`response_format`	`str \| NotGiven`	The format of the transcript output, in one of these options: `json`, `text`, `srt`, `verbose_json`, or `vtt`.	`NOT_GIVEN`
`temperature`	`float \| NotGiven`	The sampling temperature, between 0 and 1. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. If set to 0, the model will use log probability to automatically increase the temperature until certain thresholds are hit.	`NOT_GIVEN`
`extra_headers`	`Headers \| None`	Send extra headers	`None`
`extra_query`	`Query \| None`	Add additional query parameters to the request	`None`
`extra_body`	`Body \| None`	Add additional JSON properties to the request	`None`
`timeout`	`float \| Timeout \| None \| NotGiven`	Override the client-level default timeout for this request, in seconds	`NOT_GIVEN`

with_raw_response

with_raw_response() -> AsyncTranslationsWithRawResponse

with_streaming_response

with_streaming_response() -> (
    AsyncTranslationsWithStreamingResponse
)

AsyncTranslationsWithRawResponse

AsyncTranslationsWithRawResponse(
    translations: AsyncTranslations,
)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = async_to_raw_response_wrapper(create)

AsyncTranslationsWithStreamingResponse

AsyncTranslationsWithStreamingResponse(
    translations: AsyncTranslations,
)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = async_to_streamed_response_wrapper(create)

Audio

Audio(client: OpenAI)

Methods:

Name	Description
`speech`
`transcriptions`
`translations`
`with_raw_response`
`with_streaming_response`

speech

speech() -> Speech

transcriptions

transcriptions() -> Transcriptions

translations

translations() -> Translations

with_raw_response

with_raw_response() -> AudioWithRawResponse

with_streaming_response

with_streaming_response() -> AudioWithStreamingResponse

AudioWithRawResponse

AudioWithRawResponse(audio: Audio)

Methods:

Name	Description
`speech`
`transcriptions`
`translations`

speech

speech() -> SpeechWithRawResponse

transcriptions

transcriptions() -> TranscriptionsWithRawResponse

translations

translations() -> TranslationsWithRawResponse

AudioWithStreamingResponse

AudioWithStreamingResponse(audio: Audio)

Methods:

Name	Description
`speech`
`transcriptions`
`translations`

speech

speech() -> SpeechWithStreamingResponse

transcriptions

transcriptions() -> TranscriptionsWithStreamingResponse

translations

translations() -> TranslationsWithStreamingResponse

Speech

Speech(client: OpenAI)

Methods:

Name	Description
`create`	Generates audio from the input text.
`with_raw_response`
`with_streaming_response`

create

create(
    *,
    input: str,
    model: Union[str, Literal["tts-1", "tts-1-hd"]],
    voice: Literal[
        "alloy", "echo", "fable", "onyx", "nova", "shimmer"
    ],
    response_format: (
        Literal["mp3", "opus", "aac", "flac", "wav", "pcm"]
        | NotGiven
    ) = NOT_GIVEN,
    speed: float | NotGiven = NOT_GIVEN,
    extra_headers: Headers | None = None,
    extra_query: Query | None = None,
    extra_body: Body | None = None,
    timeout: float | Timeout | None | NotGiven = NOT_GIVEN
) -> HttpxBinaryResponseContent

Generates audio from the input text.

Parameters:

Name	Type	Description	Default
`input`	`str`	The text to generate audio for. The maximum length is 4096 characters.	required
`model`	`Union[str, Literal['tts-1', 'tts-1-hd']]`	One of the available TTS models: `tts-1` or `tts-1-hd`	required
`voice`	`Literal['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer']`	The voice to use when generating the audio. Supported voices are `alloy`, `echo`, `fable`, `onyx`, `nova`, and `shimmer`. Previews of the voices are available in the Text to speech guide.	required
`response_format`	`Literal['mp3', 'opus', 'aac', 'flac', 'wav', 'pcm'] \| NotGiven`	The format to audio in. Supported formats are `mp3`, `opus`, `aac`, `flac`, `wav`, and `pcm`.	`NOT_GIVEN`
`speed`	`float \| NotGiven`	The speed of the generated audio. Select a value from `0.25` to `4.0`. `1.0` is the default.	`NOT_GIVEN`
`extra_headers`	`Headers \| None`	Send extra headers	`None`
`extra_query`	`Query \| None`	Add additional query parameters to the request	`None`
`extra_body`	`Body \| None`	Add additional JSON properties to the request	`None`
`timeout`	`float \| Timeout \| None \| NotGiven`	Override the client-level default timeout for this request, in seconds	`NOT_GIVEN`

with_raw_response

with_raw_response() -> SpeechWithRawResponse

with_streaming_response

with_streaming_response() -> SpeechWithStreamingResponse

SpeechWithRawResponse

SpeechWithRawResponse(speech: Speech)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = to_raw_response_wrapper(create)

SpeechWithStreamingResponse

SpeechWithStreamingResponse(speech: Speech)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = to_custom_streamed_response_wrapper(
    create, StreamedBinaryAPIResponse
)

Transcriptions

Transcriptions(client: OpenAI)

Methods:

Name	Description
`create`	Transcribes audio into the input language.
`with_raw_response`
`with_streaming_response`

create

create(
    *,
    file: FileTypes,
    model: Union[str, Literal["whisper-1"]],
    language: str | NotGiven = NOT_GIVEN,
    prompt: str | NotGiven = NOT_GIVEN,
    response_format: (
        Literal[
            "json", "text", "srt", "verbose_json", "vtt"
        ]
        | NotGiven
    ) = NOT_GIVEN,
    temperature: float | NotGiven = NOT_GIVEN,
    timestamp_granularities: (
        List[Literal["word", "segment"]] | NotGiven
    ) = NOT_GIVEN,
    extra_headers: Headers | None = None,
    extra_query: Query | None = None,
    extra_body: Body | None = None,
    timeout: float | Timeout | None | NotGiven = NOT_GIVEN
) -> Transcription

Transcribes audio into the input language.

Parameters:

Name	Type	Description	Default
`file`	`FileTypes`	The audio file object (not file name) to transcribe, in one of these formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, or webm.	required
`model`	`Union[str, Literal['whisper-1']]`	ID of the model to use. Only `whisper-1` (which is powered by our open source Whisper V2 model) is currently available.	required
`language`	`str \| NotGiven`	The language of the input audio. Supplying the input language in ISO-639-1 format will improve accuracy and latency.	`NOT_GIVEN`
`prompt`	`str \| NotGiven`	An optional text to guide the model's style or continue a previous audio segment. The prompt should match the audio language.	`NOT_GIVEN`
`response_format`	`Literal['json', 'text', 'srt', 'verbose_json', 'vtt'] \| NotGiven`	The format of the transcript output, in one of these options: `json`, `text`, `srt`, `verbose_json`, or `vtt`.	`NOT_GIVEN`
`temperature`	`float \| NotGiven`	The sampling temperature, between 0 and 1. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. If set to 0, the model will use log probability to automatically increase the temperature until certain thresholds are hit.	`NOT_GIVEN`
`timestamp_granularities`	`List[Literal['word', 'segment']] \| NotGiven`	The timestamp granularities to populate for this transcription. `response_format` must be set `verbose_json` to use timestamp granularities. Either or both of these options are supported: `word`, or `segment`. Note: There is no additional latency for segment timestamps, but generating word timestamps incurs additional latency.	`NOT_GIVEN`
`extra_headers`	`Headers \| None`	Send extra headers	`None`
`extra_query`	`Query \| None`	Add additional query parameters to the request	`None`
`extra_body`	`Body \| None`	Add additional JSON properties to the request	`None`
`timeout`	`float \| Timeout \| None \| NotGiven`	Override the client-level default timeout for this request, in seconds	`NOT_GIVEN`

with_raw_response

with_raw_response() -> TranscriptionsWithRawResponse

with_streaming_response

with_streaming_response() -> (
    TranscriptionsWithStreamingResponse
)

TranscriptionsWithRawResponse

TranscriptionsWithRawResponse(
    transcriptions: Transcriptions,
)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = to_raw_response_wrapper(create)

TranscriptionsWithStreamingResponse

TranscriptionsWithStreamingResponse(
    transcriptions: Transcriptions,
)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = to_streamed_response_wrapper(create)

Translations

Translations(client: OpenAI)

Methods:

Name	Description
`create`	Translates audio into English.
`with_raw_response`
`with_streaming_response`

create

create(
    *,
    file: FileTypes,
    model: Union[str, Literal["whisper-1"]],
    prompt: str | NotGiven = NOT_GIVEN,
    response_format: str | NotGiven = NOT_GIVEN,
    temperature: float | NotGiven = NOT_GIVEN,
    extra_headers: Headers | None = None,
    extra_query: Query | None = None,
    extra_body: Body | None = None,
    timeout: float | Timeout | None | NotGiven = NOT_GIVEN
) -> Translation

Translates audio into English.

Parameters:

Name	Type	Description	Default
`file`	`FileTypes`	The audio file object (not file name) translate, in one of these formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, or webm.	required
`model`	`Union[str, Literal['whisper-1']]`	ID of the model to use. Only `whisper-1` (which is powered by our open source Whisper V2 model) is currently available.	required
`prompt`	`str \| NotGiven`	An optional text to guide the model's style or continue a previous audio segment. The prompt should be in English.	`NOT_GIVEN`
`response_format`	`str \| NotGiven`	The format of the transcript output, in one of these options: `json`, `text`, `srt`, `verbose_json`, or `vtt`.	`NOT_GIVEN`
`temperature`	`float \| NotGiven`	The sampling temperature, between 0 and 1. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. If set to 0, the model will use log probability to automatically increase the temperature until certain thresholds are hit.	`NOT_GIVEN`
`extra_headers`	`Headers \| None`	Send extra headers	`None`
`extra_query`	`Query \| None`	Add additional query parameters to the request	`None`
`extra_body`	`Body \| None`	Add additional JSON properties to the request	`None`
`timeout`	`float \| Timeout \| None \| NotGiven`	Override the client-level default timeout for this request, in seconds	`NOT_GIVEN`

with_raw_response

with_raw_response() -> TranslationsWithRawResponse

with_streaming_response

with_streaming_response() -> (
    TranslationsWithStreamingResponse
)

TranslationsWithRawResponse

TranslationsWithRawResponse(translations: Translations)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = to_raw_response_wrapper(create)

TranslationsWithStreamingResponse

TranslationsWithStreamingResponse(
    translations: Translations,
)

Attributes:

Name	Type	Description
`create`

create `instance-attribute`

create = to_streamed_response_wrapper(create)

openai.resources.audio

AsyncAudio

speech

transcriptions

translations

with_raw_response

with_streaming_response

AsyncAudioWithRawResponse

speech

transcriptions

translations

AsyncAudioWithStreamingResponse

speech

transcriptions

translations

AsyncSpeech

create async

with_raw_response

with_streaming_response

AsyncSpeechWithRawResponse

create instance-attribute

AsyncSpeechWithStreamingResponse

create instance-attribute

AsyncTranscriptions

create async

with_raw_response

with_streaming_response

AsyncTranscriptionsWithRawResponse

create instance-attribute

AsyncTranscriptionsWithStreamingResponse

create instance-attribute

AsyncTranslations

create async

with_raw_response

with_streaming_response

AsyncTranslationsWithRawResponse

create instance-attribute

AsyncTranslationsWithStreamingResponse

create instance-attribute

Audio

speech

transcriptions

translations

with_raw_response

with_streaming_response

AudioWithRawResponse

speech

transcriptions

translations

AudioWithStreamingResponse

speech

transcriptions

translations

Speech

create

with_raw_response

with_streaming_response

SpeechWithRawResponse

create instance-attribute

SpeechWithStreamingResponse

create instance-attribute

Transcriptions

create

with_raw_response

with_streaming_response

TranscriptionsWithRawResponse

create instance-attribute

TranscriptionsWithStreamingResponse

create instance-attribute

Translations

create

with_raw_response

with_streaming_response

TranslationsWithRawResponse

create instance-attribute

TranslationsWithStreamingResponse

create instance-attribute

create `async`

create `instance-attribute`

create `instance-attribute`

create `async`

create `instance-attribute`

create `instance-attribute`

create `async`

create `instance-attribute`

create `instance-attribute`

create `instance-attribute`

create `instance-attribute`

create `instance-attribute`

create `instance-attribute`

create `instance-attribute`

create `instance-attribute`