AudioTranscription

data class AudioTranscription(val text: String, val language: String? = null, val duration: Double? = null, val segments: List<Segment>? = null, val words: List<Word>? = null, val logprobs: List<AudioLogprob>? = null, val languages: List<TranscriptionLanguage>? = null, val usage: AudioUsage? = null, val diarizedSegments: List<DiarizedSegment>? = null, val task: String? = null)

Modern transcription result with model-specific confidence, usage and speaker details.

Timestamp and speaker-labeled segments are represented separately.

Constructors

Link copied to clipboard
constructor(text: String, language: String? = null, duration: Double? = null, segments: List<Segment>? = null, words: List<Word>? = null, logprobs: List<AudioLogprob>? = null, languages: List<TranscriptionLanguage>? = null, usage: AudioUsage? = null, diarizedSegments: List<DiarizedSegment>? = null, task: String? = null)

Properties

Link copied to clipboard

Speaker-labeled segments for diarized JSON responses.

Link copied to clipboard
@SerialName(value = "duration")
val duration: Double? = null
Link copied to clipboard
@SerialName(value = "language")
val language: String? = null
Link copied to clipboard

Languages detected by multilingual models.

Link copied to clipboard
val logprobs: List<AudioLogprob>? = null

Token confidence information when requested.

Link copied to clipboard
@SerialName(value = "segments")
val segments: List<Segment>? = null
Link copied to clipboard
val task: String? = null

Task name returned by diarized responses.

Link copied to clipboard
@SerialName(value = "text")
val text: String

Transcription text. The format depends on TranscriptionRequest's responseFormat.

Link copied to clipboard
val usage: AudioUsage? = null

Token-based or duration-based request usage.

Link copied to clipboard
@SerialName(value = "words")
val words: List<Word>? = null