Export Transcripts: SRT, VTT, JSON, Markdown & DOCX

Once a transcription is complete, export it in any format with a single request. Six formats cover subtitle workflows, downstream processing, and content pipelines.

Endpoint

bash

GET /api/v1/transcriptions/{id}/export?format={format}
Scope: exports:read

Supported formats

Format	MIME type	Use case
txt	text/plain	Raw transcript text, no timestamps
srt	text/plain	SubRip subtitles — YouTube, Premiere, DaVinci
vtt	text/vtt	WebVTT — HTML5 video, Notion, Loom
json	application/json	Timestamped words + segments for custom processing
md	text/markdown	Markdown with speaker labels and timestamps
docx	application/vnd.openxmlformats-officedocument...	Word document for editorial workflows

Export formats panel — TXT, SRT, DOCX, MD, VTT, and Advanced Export — All export formats available from the transcription detail page

SRT example

const srt = await client.transcriptions.export(99, { format: 'srt' });
// "1\n00:00:00,000 --> 00:00:03,240\nNever gonna give you up\n\n2\n..."

JSON (timestamped words)

The json format returns segments with word-level timestamps:

json

{
  "transcriptionId": 99,
  "language": "en",
  "durationSeconds": 212,
  "segments": [
    {
      "id": 0,
      "start": 0.0,
      "end": 3.24,
      "text": "Never gonna give you up",
      "speaker": "Speaker 1",
      "words": [
        {
          "word": "Never",
          "start": 0.0,
          "end": 0.4,
          "probability": 0.99
        },
        {
          "word": "gonna",
          "start": 0.4,
          "end": 0.72,
          "probability": 0.98
        }
      ]
    }
  ]
}

SDK usage

// Any format — returns raw string content
const srt = await client.transcriptions.export(99, { format: 'srt' });

// JSON format — returns parsed object with segments array
const json = await client.transcriptions.export(99, { format: 'json' });

Notes

Speaker labels in json and md are only available on plans that include speaker diarization (Pro and above). Free-tier transcripts omit speaker fields.
docx includes a cover page with the source URL, language, and duration.
Exports are generated on demand — not pre-stored. Each call re-renders the format from the stored segment data.

FAQ

Can I export multiple formats from the same transcript?

Yes. Each export call is independent — you can call the endpoint as many times as you need, with different format values. There's no extra charge per export call.

Do SRT and VTT files include speaker names?

No. SRT and VTT only contain timecodes and text. Speaker labels are available in the json and md formats on Pro plans and above.

How are word-level timestamps structured in the JSON export?

The json export returns a segments array. Each segment has start, end, text, and a nested words array with per-word start, end, and confidence scores.