Fast Audio & Video Transcriber

Turn a direct audio or video file URL into multilingual text, timestamped segments, SRT, and WebVTT. This is a simple speech-to-text / transcription API for automation: GPU-powered Whisper transcription, automatic language detection, and JSON output. No external transcription API key is required.

What it does

Give the Actor a direct downloadable HTTPS media URL. It returns:

• a full transcript (audio to text / video to text)
• detected language
• timestamped segments
• an SRT subtitle track
• a WebVTT subtitle track

Video files are transcribed from their audio track. This version does not scrape YouTube, TikTok, Instagram, or other HTML/player pages.

Pricing

Pay per event: $0.012 per started audio minute.

charged_minutes = ceil(audio_duration_seconds / 60)

Duration	Charged minutes	Price
10 minutes	10	$0.12
30 minutes	30	$0.36
1 hour / 60 minutes	60	$0.72
2 hours / 120 minutes	120	$1.44
1,000 minutes	1,000	$12.00

There is no charge for unsuccessful transcription events. No per-file fee, no subtitle fee, and no language-detection fee. Apify platform usage is included in this event price.

Input

mediaUrl must be a publicly downloadable HTTPS URL that points directly to an audio or video file (for example a public S3 object, a presigned HTTPS URL, a CDN file, a podcast MP3, or an MP4).

{
  "mediaUrl": "https://example.com/audio.mp3"
}

Optional:

{
  "mediaUrl": "https://example.com/audio.mp3",
  "language": "en",
  "maxAudioMinutes": 120
}

Leave language empty for automatic multilingual detection. maxAudioMinutes defaults to 120 and cannot exceed 120.

Supported formats

Audio: mp3, wav, m4a, aac, ogg, flac
Video: mp4, mov, webm, mkv

Actual support depends on media decoding.

Output

Each successful run stores one Dataset item and the same object in Actor OUTPUT:

{
  "id": "c2f1a0b4-9c3e-4d21-9a10-4b2c8e7f1d90",
  "source": { "url": "https://example.com/audio.mp3" },
  "model": "whisper-turbo",
  "language": "en",
  "languageProbability": 0.99,
  "durationSeconds": 61.2,
  "chargedMinutes": 2,
  "processingSeconds": 1.4,
  "text": "…",
  "segments": [{ "start": 0.0, "end": 2.4, "text": "…" }],
  "srt": "1\n00:00:00,000 --> 00:00:02,400\n…\n",
  "vtt": "WEBVTT\n\n00:00:00.000 --> 00:00:02.400\n…\n"
}

Use the text, timestamps, segments, SRT, and WebVTT together. Signed URL query strings are stripped from source.url.

Use cases

Podcasts, interviews, meetings, lectures, videos, subtitle generator workflows, searchable archives, RAG / LLM pipelines, content analysis, and media automation.

API usage

Call the Actor through Apify. Never put a real token in source control.

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("ozwstone/transcriber-adapter").call(
    run_input={"mediaUrl": "https://example.com/audio.mp3"}
)
item = client.dataset(run["defaultDatasetId"]).list_items().items[0]
print(item["text"])

curl -X POST "https://api.apify.com/v2/acts/ozwstone~transcriber-adapter/runs?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"mediaUrl":"https://example.com/audio.mp3"}'

Limits

• Direct media URLs only
• HTTPS only
• Maximum 120 audio minutes per run
• No YouTube / TikTok / Instagram page URLs yet
• No speaker diarization yet
• No translation yet
• No file uploads through this Actor

Privacy & security

Media is processed for transcription. Backend temporary files are deleted after the job. Signed URL query strings and fragments are not written to results or logs. You must have the right to process submitted media. This README does not claim GDPR, HIPAA, or other certifications.