Skip to main content

​
Overview

The Segment and Token types represent transcribed audio with timing information. A Segment represents a complete phrase or sentence, while Token represents individual words within that segment.

​
Segment

Represents a segment of transcribed audio with timing information. May include detailed token-level (word-by-word) information when available.

​
Type definition

type Segment = Token & {
    tokens?: Token[];
};

​
Properties

​
text
string
required
The transcribed text for this segment.
​
start
number
required
Start time of the segment in seconds.
​
end
number
required
End time of the segment in seconds.
​
confidence
number
Confidence score for this transcription (between 0 and 1).Higher values indicate greater confidence in the transcription accuracy.
​
tokens
Token[]
Word-by-word breakdown of the transcription with individual timings.

​
Token

Represents a token (word or phrase) in the transcription with timing information.

​
Type definition

type Token = TimeRange & {
    confidence?: number;
    text: string;
};

​
Properties

​
text
string
required
The transcribed text (word or phrase).
​
start
number
required
Start time in seconds.Inherited from TimeRange type (from ffmpeg-simplified).
​
end
number
required
End time in seconds.Inherited from TimeRange type (from ffmpeg-simplified).
​
confidence
number
Confidence score for this transcription (between 0 and 1).

​
Usage examples

import { transcribe } from 'tafrigh';

const segments = await transcribe('audio.mp3');

// Access segment data
segments.forEach((segment) => {
  console.log(`[${segment.start}s - ${segment.end}s]: ${segment.text}`);
  console.log(`Confidence: ${segment.confidence}`);
});

​
Response structure example

[
  {
    "text": "Hello world",
    "start": 0,
    "end": 2.5,
    "confidence": 0.95,
    "tokens": [
      {
        "text": "Hello",
        "start": 0,
        "end": 1.2,
        "confidence": 0.97
      },
      {
        "text": "world",
        "start": 1.3,
        "end": 2.5,
        "confidence": 0.93
      }
    ]
  },
  {
    "text": "This is a test",
    "start": 3.0,
    "end": 5.5,
    "confidence": 0.88
  }
]

transcribe

Returns an array of Segment objects

Callbacks

onTranscriptionFinished receives Segment[]

Error handling

TranscriptionError includes partial segments

Resume failed transcriptions

Continue from partial transcriptions

Build docs developers (and LLMs) love

Get started for freeTalk to us