Skip to main content
Interface representing the transcription result returned by transcribe(), transcribeData(), and realtime transcription methods.

​
Properties

​
result
string
required
The complete transcribed text. Contains the full transcription result as a single string.
​
language
string
required
The detected or specified language code (e.g., ‘en’, ‘es’, ‘fr’). If language was set to ‘auto’ in options, this will be the detected language.
​
segments
array
required
Array of transcription segments with timestamps. Each segment represents a portion of the transcribed audio.
​
isAborted
boolean
required
Whether the transcription was aborted by calling stop(). When true, the transcription was cancelled before completion.

​
Usage Example

import { initWhisper } from 'whisper.rn'

const context = await initWhisper({
  filePath: 'path/to/model.bin'
})

const { promise } = context.transcribe('path/to/audio.wav', {
  language: 'en'
})

const result: TranscribeResult = await promise

console.log('Full text:', result.result)
console.log('Language:', result.language)
console.log('Was aborted:', result.isAborted)

// Process segments with timestamps
result.segments.forEach(segment => {
  console.log(`[${segment.t0}ms - ${segment.t1}ms]: ${segment.text}`)
})

​
Example Response

{
  "result": "Hello, this is a test transcription.",
  "language": "en",
  "isAborted": false,
  "segments": [
    {
      "text": "Hello, this is a test",
      "t0": 0,
      "t1": 1500
    },
    {
      "text": " transcription.",
      "t0": 1500,
      "t1": 2800
    }
  ]
}

​
Notes

  • Timestamps are in milliseconds, not seconds
  • Segments may overlap slightly depending on the model’s processing
  • When isAborted is true, the result and segments may be partial or empty
  • The language field uses ISO 639-1 language codes

Build docs developers (and LLMs) love

Get started for freeTalk to us