Skip to contents

Audio API

Source: R/audio.R
audio.Rd

Turn audio into text or text into audio.

  • oai_create_speec(): Generates audio from the input text.

  • oai_create_transcription(): Transcribes audio into the input language.

  • oai_create_translation(): Translates audio into English.

Usage

oai_create_speech(
  input,
  path,
  model = c("tts-1", "tts-1-hd", "gpt-4o-mini-tts"),
  voice = c("alloy", "ash", "ballad", "coral", "echo", "fable", "onyx", "nova", "sage",
    "shimmer", "verse"),
  instruction = NULL,
  response_format = c("mp3", "opus", "aac", "flac", "wav", "pcm"),
  speed = NULL,
  .async = FALSE
)

oai_create_transcription(
  file,
  model = "whisper-1",
  language = NULL,
  prompt = NULL,
  response_format = c("json", "text", "srt", "verbose_json", "vtt"),
  temperature = NULL,
  timestamp_granularities = NULL,
  name = NULL,
  .classify_response = TRUE,
  .async = FALSE
)

oai_create_translation(
  file,
  model = "whisper-1",
  prompt = NULL,
  response_format = c("json", "text", "srt", "verbose_json", "vtt"),
  temperature = NULL,
  name = NULL,
  .classify_response = TRUE,
  .async = FALSE
)

Arguments

input

Character. The text to generate audio for. The maximum length is 4096 characters.

path

Character. The path to save the audio file.

model

Character. ID of the model to use. Only whisper-1 (which is powered by our open source Whisper V2 model) is currently available.

voice

Character. The voice to use when generating the audio. Supported voices are alloy, ash, coral, echo, fable, onyx, nova, sage and shimmer. Previews of the voices are available in the Text to speech guide.

instruction

Character. Control the voice of your generated audio with additional instructions. Does not work with tts-1 or tts-1-hd.

response_format

Character. The format of the output, in one of these options: json, text, srt, verbose_json, or vtt.

speed

Numeric. The speed of the generated audio. Select a value from 0.25 to 4.0. 1.0 is the default.

.async

Logical. If TRUE, the request is performed asynchronously.

file

Character. The audio file object (not file name) translate, in one of these formats: flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, or webm.

language

Character. The language of the input audio. Supplying the input language in ISO-639-1 format will improve accuracy and latency.

prompt

Character. An optional text to guide the model's style or continue a previous audio segment. The prompt should be in English.

temperature

Numeric. 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.

timestamp_granularities

List. 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.

name

Character. The file name to use for the upload.

.classify_response

Logical. If TRUE (default), the response is classified as an R6 object. If FALSE, the response is returned as a list.

Value

Returns the audio file content.

Returns the transcription object or a verbose transcription object.

Returns the translated text.