Audio & Video Transcriber: Whisper Speech to Text, Subtitles & Podcast Transcripts

Turn audio and video files into accurate text with OpenAI Whisper (open-source model), in about 100 languages. Paste direct links to MP3, M4A, WAV, MP4 or other media files, or a podcast RSS feed. You get the full transcript, timestamped segments and ready-to-use SRT and WebVTT subtitles, for $0.008 per minute. No API keys or subscriptions.

What you get

• Full transcript as clean text, with paragraphs where the speaker pauses.
• SRT and WebVTT subtitles, ready for video players, YouTube uploads or editing tools.
• Timestamped segments (JSON) with start and end times in seconds.
• Detected language and confidence, or translation to English with one switch.
• Podcast support: give an RSS feed and the latest episodes are transcribed automatically.

Typical uses:

• Podcasters and creators: show notes, blog posts, captions and searchable archives.
• Marketers and SEO: turn webinars and videos into articles and quotes.
• Researchers and journalists: transcribe interviews, lectures and meetings.
• AI builders: feed transcripts to LLMs, RAG pipelines and summarization flows.

Input

• Audio or video file URLs: direct links to media files, up to 4 hours each. Most formats work: mp3, m4a, aac, wav, flac, ogg, opus, mp4, mov, mkv and webm.
• Podcast RSS feeds: the latest N episodes of each feed.
• Spoken language: auto-detect by default.
• Translate to English: returns an English version instead of the original language.
• Max minutes per file: caps the length transcribed and charged.

Not supported: YouTube, TikTok, Instagram, Facebook, X, Vimeo and Spotify page links, because those platforms don't allow downloading. Use a direct link to the media file you own or have permission to use.

Output example

{
    "title": null,
    "url": "https://archive.org/download/gettysburg_johng_librivox/gettysburg_address.mp3",
    "source": "url",
    "durationSec": 157.89,
    "transcribedSec": 157.89,
    "truncated": false,
    "language": "en",
    "languageProbability": 1,
    "wordCount": 298,
    "text": "This is a LibriVox recording. All LibriVox recordings are in the public domain. ... Four score and seven years ago, our fathers brought forth on this continent a new nation...",
    "srt": "1\n00:00:00,050 --> 00:00:05,750\nThis is a LibriVox recording. All LibriVox recordings are in the public domain.\n...",
    "vtt": "WEBVTT\n\n00:00:00.050 --> 00:00:05.750\nThis is a LibriVox recording...",
    "segments": [{ "start": 0.05, "end": 5.75, "text": "This is a LibriVox recording. All LibriVox recordings are in the public domain." }],
    "billedMinutes": 3,
    "error": null
}

Real output for the public-domain LibriVox recording above (texts shortened). Files that cannot be downloaded or have no audio get an error message and are not charged.

How much does it cost?

Event	Price (free plan)
Run start	$0.02
Audio minute (rounded up per file)	$0.008

Paid Apify plans get lower per-minute prices. Examples on the free plan:

• A 30-minute podcast episode: $0.26
• 10 one-hour lectures: $4.82

Set Maximum cost per run to cap spending. The run stops before starting a file it can't afford.

FAQ

Which model is used? OpenAI's open-source Whisper "small" model, running on dedicated servers with voice-activity detection to skip silence. It works well for podcasts, interviews and lectures. Very noisy audio or heavy accents reduce accuracy, as with any speech recognition.

How long does it take? Usually about a fifth of the audio length once the service is warm, plus up to a minute to start. Long files are split into 5-minute parts that are processed in parallel.

Is my audio stored? Files are downloaded to temporary storage only for processing and deleted automatically within 24 hours. Transcripts are saved only in your Apify dataset.

Can I use it from code or an AI agent? Yes. Use the Apify API, the JavaScript or Python client, Make, Zapier, n8n, or the Apify MCP server.

Support

Open an issue in the Issues tab. Issues are reviewed regularly.