Transcribe a video

Transcription uses AI to generate a text track from a video’s audio. The generated text track is added to the video as an auto-generated subtitle track and is also available as a plain-text transcript.

AI features require a plan with AI features enabled. The API token that you use must have the update permission.

Start transcription

Start a transcription by sending a POST request to the /videos/:id/transcribe endpoint. The language field tells the transcription engine which language the audio is in.

CURL


curl -X POST https://app.ignitevideo.cloud/api/videos/[VIDEO_ID]/transcribe \
-H "Authorization: Bearer YOUR_TOKEN" \
-H "Content-Type: application/json" \
-d '{"language": "en"}'

JavaScript / Node.js


const data = {
language: "en" // BCP-47 language tag, e.g. "en", "en-US", "de", "de-DE"
};
 
const response = await fetch('https://app.ignitevideo.cloud/api/videos/[VIDEO_ID]/transcribe', {
method: 'POST',
body: JSON.stringify(data),
headers: {
  'Authorization': 'Bearer YOUR_TOKEN',
  'Content-Type': 'application/json'
}
});
 
const result = await response.json();

PHP


$client = new GuzzleHttp\Client();
$data = [
"language" => "en" // BCP-47 language tag, e.g. "en", "en-US", "de", "de-DE"
];
 
$response = $client->request('POST', 'https://app.ignitevideo.cloud/api/videos/[VIDEO_ID]/transcribe', [
'body' => json_encode($data),
'headers' => [
  'Authorization' => 'Bearer YOUR_TOKEN',
  'Content-Type' => 'application/json'
]
]);
 
$result = json_decode($response->getBody(), true);

Parameters

Field	Required	Description
`language`	Yes	Language of the audio as a BCP-47 tag (`en`, `en-US`, `de`, `de-DE`, …).

Response

Transcription runs asynchronously. The request returns immediately with one or more transcription job records. Videos longer than 3 hours are split into multiple chunks, so the response can contain more than one entry.


[
  {
    "id": string,
    "status": "QUEUED" | "IN_PROGRESS" | "COMPLETED" | "FAILED" | "NONE",
    "language": string,
    "jobName": string
  }
]

When a job completes, an auto-generated text track is added to the video and the transcript is filled in. Poll the get video endpoint and read texttracks and transcriptions (or autoTranscription) to retrieve the result.

Transcribe automatically on upload

You can also start transcription automatically once encoding finishes. Set autoTranscribe: true and a valid language when you create a video. This requires the create permission instead of update.


{
  "title": "Your video title",
  "language": "en",
  "autoTranscribe": true
}