Audio & Video to Text — Whisper Transcriber

Turn audio and video into clean text and ready-to-use subtitles—without an API key.

Upload a file, paste a direct media link, or connect a file produced by another Apify Actor. This Actor runs the multilingual Whisper large-v3-turbo model through fast native whisper inference and returns both a readable transcript and timestamped SRT subtitles.

What you get

• Plain-text transcript, ready for notes, search, summaries, or RAG workflows
• Timestamped SRT subtitles for video editors and publishing platforms
• Automatic language detection or an optional language hint
• Optional translation to English
• Support for audio and video formats handled by FFmpeg, including MP3, WAV, M4A, MP4, MOV, WebM, and more
• A single 30-second processing window in memory at a time—well suited to long recordings
• One model load per Actor run for efficient long-form transcription

Three ways to add media

1. Paste a direct media URL

Use a public or signed link to an actual audio or video file:

{
  "audioUrl": "https://example.com/interview.mp3",
  "language": "cs",
  "translateToEnglish": false
}

2. Upload a file

In the Apify input form, use Upload media or media URL to select an audio or video file from your computer. Apify stores it for the run and passes its file URL to the Actor—no separate hosting needed.

3. Use media from another Actor

If another Actor saved media in an Apify Key-value Store, select that store in Source key-value store and enter the file's record key:

{
  "sourceKeyValueStore": "YOUR_STORE_ID",
  "sourceFileKey": "meeting-recording.mp4",
  "language": "en"
}

The Actor requests read-only access only to the storage you select.

Provide exactly one source: audioUrl, uploaded media, or the sourceKeyValueStore + sourceFileKey pair.

Results

Every successful run writes one item to the default Dataset:

{
  "audioUrl": "https://example.com/interview.mp3",
  "durationSeconds": 125.4,
  "billedMinutes": 3,
  "language": "cs",
  "translatedToEnglish": false,
  "model": "base",
  "transcript": "Dobrý den. Vítejte v dnešním rozhovoru.",
  "subtitlesSrt": "1\n00:00:00,000 --> 00:00:02,400\nDobrý den.\n"
}

Built for long recordings

Long media is processed in 30-second chunks by default. Adjacent chunks overlap by one second to protect speech at boundaries, and duplicate subtitle lines are removed. Set chunkSeconds to a whole number from 5 to 30 when needed.

If Apify migrates a long-running Actor to another server, the Actor saves its completed chunks in the run's default Key-value Store and resumes from the next chunk after restart. The temporary media download is repeated, but already transcribed audio is not processed again.

Pricing

The Actor charges one audio-minute event for every started minute only after a transcript has been produced successfully. Check the Apify pricing panel for the current event price and applicable platform-usage charges before starting a run.

Important notes

• Use a direct media-file URL, not a web page that merely contains an embedded player. YouTube, TikTok, and Instagram page URLs are not currently supported.
• You are responsible for having the rights and consent needed to process the submitted media.
• Uploaded files and Dataset results follow the storage and retention settings of the Apify account that starts the run.