Audio & Video Transcriber – Whisper Speech to Text, SRT

Transcribe audio and video files (MP3, WAV, M4A, FLAC, OGG, MP4, MOV, WebM...) and podcast episodes to text with OpenAI Whisper (faster-whisper) running inside the Actor. You get the full transcript, timestamped segments, optional word timings, auto-detected language (99 languages), translation to English, and ready-made SRT / VTT subtitle files. No OpenAI key is needed. Pricing is per audio minute and starts at $0.005.

Why use it

• Cheap. $0.005/min (standard) is cheaper than the Whisper API ($0.006/min), with no API key and no 25 MB upload limit.
• Podcasts built in. Paste an RSS feed and it transcribes the latest N episodes automatically.
• Long files. Multi-hour recordings are decoded in chunks. Big video files work too, since only the audio track is used.
• Subtitles out of the box. .srt, .vtt and .txt files are saved for every file.
• Two quality levels. Standard (Whisper base) is fast and cheap. Accurate (Whisper small) gives better punctuation, names and non-English results.
• Fair billing. You pay per started minute actually transcribed. Silent or no-speech files and failed downloads are free.

How to use it

1. Add direct links to audio or video files under Audio/video URLs: S3, Dropbox, Google Drive (public link), your CDN, or podcast MP3s. You can also add Podcast RSS feeds.
2. Choose Quality and Language (auto-detect by default), and optionally Translate to English or Word timestamps.
3. Click Start. Get text and segments from the dataset, or download the subtitles from srtUrl / vttUrl.

Note: video-platform page links (YouTube, TikTok, Instagram, Vimeo, X...) aren't supported. Use a direct media file you have the right to process.

Input example

{
  "sources": [{ "url": "https://raw.githubusercontent.com/openai/whisper/main/tests/jfk.flac" }],
  "podcastFeeds": [{ "url": "https://feeds.simplecast.com/54nAGcIl" }],
  "maxEpisodesPerFeed": 1,
  "quality": "standard",
  "language": "auto"
}

Output example

{
  "url": "https://raw.githubusercontent.com/openai/whisper/main/tests/jfk.flac",
  "status": "OK",
  "model": "whisper-base",
  "durationSeconds": 11.0,
  "language": "en",
  "languageProbability": 0.925,
  "text": "And so my fellow Americans ask not what your country can do for you Ask what you can do for your country",
  "billedMinutes": 1,
  "wordCount": 22,
  "segments": [
    { "start": 0.0, "end": 7.28, "text": "And so my fellow Americans ask not what your country can do for you" },
    { "start": 7.98, "end": 9.98, "text": "Ask what you can do for your country" }
  ],
  "srtUrl": "https://api.apify.com/v2/key-value-stores/…/records/jfk-1.srt",
  "vttUrl": "https://api.apify.com/v2/key-value-stores/…/records/jfk-1.vtt",
  "txtUrl": "https://api.apify.com/v2/key-value-stores/…/records/jfk-1.txt"
}

status values: OK, NO_SPEECH (free), FAILED, HTTP_ERROR, UNSUPPORTED_SOURCE, SKIPPED_ROBOTS_TXT.

Pricing (pay per event)

Event	Price
Audio minute transcribed (Standard, Whisper base)	$0.005
Audio minute transcribed (Accurate, Whisper small)	$0.012

Billing is per started minute of audio transcribed, per file. Example: a 45-minute podcast costs $0.225 on Standard or $0.54 on Accurate. Use Max duration per file to cap long files. Only the first N minutes are transcribed and billed.

FAQ

How fast is it? Standard transcribes roughly 8–10x faster than real time on the default 4 GB memory. A 1-hour episode takes about 7 minutes. Accurate is about 3x slower.

Which languages? All 99 Whisper languages, with automatic detection. Set the Language code (for example de or es) to skip detection on short clips.

Speaker labels (diarization)? Not yet. You get timestamps per segment and optionally per word.

Can I transcribe YouTube videos? Not through this Actor. Downloading from video platforms breaks their terms. Use your own files or podcast feeds.

Is my audio stored? No. Only the transcript and subtitle files are saved in your own run storage.