Audio & Video Transcriber: Whisper Speech to Text, SRT

Audio & Video Transcriber turns audio and video files, podcast RSS feeds and Apple Podcasts shows into accurate text with timestamps, using OpenAI's open-source Whisper speech recognition model running inside the Actor. You get the full transcript, timed segments, optional word-level timestamps and ready-made SRT, VTT and TXT subtitle files for every file.

Paste direct file links, Dropbox, Google Drive or OneDrive share links, a podcast feed or an Apple Podcasts link. It detects 99 languages automatically and can translate any of them into English.

✅ Pay per result: from $0.0085 per audio minute ($0.51 per hour) · 🆓 Files that fail are never charged · 🔑 No API key, no OpenAI account, audio is never sent to a third-party AI service

▶ Try it with one click · 📄 Real example output · 💲 Pricing · 🔌 API & integrations · ❓ FAQ · 🛟 Support

What does Audio & Video Transcriber do?

• 🎙️ Speech to text for audio and video: MP3, M4A, WAV, OGG, OPUS, FLAC, AAC, WebM, MP4, MOV, MKV and anything else FFmpeg can read
• 📻 Whole podcasts in one run: give a podcast RSS feed or an Apple Podcasts link and it transcribes the newest episodes (or every episode since a date), with episode title, date and page link on each row
• 🔗 Share links just work: Dropbox, Google Drive and OneDrive links are converted to direct downloads for you
• 🎬 Subtitles included: SRT and VTT files with timings, plus a clean TXT transcript and an optional JSON file, each with a download link
• 🌍 99 languages, auto-detected, or translate any language straight into English text and subtitles
• ⏱️ Timestamps: segment timings on every row and optional word-level timestamps with confidence scores
• 🧠 Two models: Standard (Whisper base, fast and cheap) and High accuracy (Whisper small with beam search) for difficult audio, accents and names
• 🏷️ Vocabulary hint: give product names, people or jargon so they are spelled correctly
• 🛡️ Spending control: per-file minute cap, file size cap, and it stops cleanly at your Max cost per run

Use cases

• Podcasters and media teams: transcripts and show notes for every episode, subtitles for video clips
• Marketers and SEO teams: turn webinars, interviews and podcasts into blog posts, quotes and searchable text
• Researchers and journalists: transcribe interviews, focus groups and public speeches with timestamps for citations
• AI and RAG builders: feed spoken content into LLMs, vector databases and AI agents
• Accessibility: captions (SRT/VTT) for course videos, product demos and social clips
• Monitoring: schedule a podcast feed and get new episodes transcribed automatically

Why choose this transcriber?

	This Actor	Typical Store transcribers
Podcast RSS feeds and Apple Podcasts links	✅ newest N episodes or all since a date	Usually single files only
Dropbox, Google Drive, OneDrive share links	✅ converted automatically	Often need a direct link
SRT + VTT + TXT (+ JSON) files with download links	✅ every file	Often text only, or extra cost
Word-level timestamps, translation to English, vocabulary hint	✅ included, no surcharge	Mixed
Price	$0.0085 per minute Standard, $0.012 High accuracy	Most charge $0.006 to $0.05 per minute; some add per-file or usage fees
Failed files	Free	Varies

How to use it

1. Click Try for free. The input is already filled in with two short public-domain speeches (26 s and 1 min 44 s). Press Start: it finishes in under a minute and costs about $0.03.
2. Replace the links with your own: files, share links, podcast feeds or Apple Podcasts show or episode links. You can also upload a text file with one link per line.
3. Pick the model (Standard or High accuracy), the language (or Auto-detect) and the subtitle formats.
4. For podcasts, set Max episodes per feed or Only episodes published since.
5. Download the results as JSON, CSV or Excel, or open the SRT/VTT/TXT links in each row.

Input example

{
  "sources": [
    { "url": "https://upload.wikimedia.org/wikipedia/commons/4/46/1941_Roosevelt_speech_pearlharbor_p1.ogg" },
    { "url": "https://upload.wikimedia.org/wikipedia/commons/3/30/Wilson_-_Address_to_the_American_Indians_edit.ogg" }
  ]
}

Other options: model (standard or accurate), language (auto, en, es, de, ...), translateToEnglish, subtitleFormats (srt, vtt, txt, json), includeSegments, wordTimestamps, vocabularyHint, skipSilence, maxEpisodesPerFeed, episodesSince, maxFiles, maxMinutesPerFile, maxFileSizeMb.

Output example (from a real public run)

First row of the public example run (segments shortened with ...):

{
  "source": "https://upload.wikimedia.org/wikipedia/commons/4/46/1941_Roosevelt_speech_pearlharbor_p1.ogg",
  "mediaUrl": "https://upload.wikimedia.org/wikipedia/commons/4/46/1941_Roosevelt_speech_pearlharbor_p1.ogg",
  "title": "1941 Roosevelt speech pearlharbor p1",
  "status": "ok",
  "language": "en",
  "languageProbability": 0.983,
  "task": "transcribe",
  "model": "whisper-base",
  "durationSeconds": 26.1,
  "transcribedSeconds": 26.1,
  "truncated": false,
  "minutesCharged": 1,
  "fileSizeMb": 0.36,
  "wordCount": 30,
  "text": "Yesterday, December 7, 1941, date which will live in infamy, at its states of America, was suddenly and deliberately attacked by naval and air forces of the Empire of Japan.",
  "files": {
    "srt": "https://api.apify.com/v2/key-value-stores/ZdQQE7BD9dUkftgUO/records/001-1941-roosevelt-speech-pearlharbor-p1.srt?signature=XCYLk4v1RdNKFghYQE2j",
    "vtt": "https://api.apify.com/v2/key-value-stores/ZdQQE7BD9dUkftgUO/records/001-1941-roosevelt-speech-pearlharbor-p1.vtt?signature=cp3OUD1AHYBBruqrAmGO",
    "txt": "https://api.apify.com/v2/key-value-stores/ZdQQE7BD9dUkftgUO/records/001-1941-roosevelt-speech-pearlharbor-p1.txt?signature=KDSJfTKrsVWNKksCSiVe"
  },
  "processingSeconds": 2.4,
  "segments": [
    { "id": 1, "start": 0.24, "end": 13.26, "text": "Yesterday, December 7, 1941, date which will live in infamy," },
    ...
  ]
}

Full results: JSON · CSV · SRT file. This 1941 recording is noisy; the High accuracy model handles hard audio better.

Podcast rows also carry podcastTitle, episodeTitle, episodePublishedAt, episodePageUrl, episodeGuid and feedUrl. With wordTimestamps on, each segment has a words list with start, end, word and probability.

Field	Meaning
status	ok or failed (failed rows have an error and are free)
text	Full transcript
segments	Timed segments (start, end in seconds, text)
language, languageProbability	Detected language and confidence
durationSeconds, transcribedSeconds, truncated	File length, how much was transcribed, and whether your minute cap cut it short
minutesCharged	Started minutes billed for this file
files	Download links for the SRT, VTT, TXT and JSON files

How much does it cost to transcribe audio and video?

Pay-per-event pricing: you pay per started minute of audio actually transcribed. No platform usage charges on top.

Event	Price	Notes
Audio minute, Standard (Whisper base)	$0.0085	Silver $0.0075, Gold and above $0.0065
Audio minute, High accuracy (Whisper small, beam search)	$0.012	Silver $0.0105, Gold and above $0.009
Run start	$0.0005 per GB of memory	Default 4 GB = $0.002 per run

Minutes are counted per file and rounded up (a 26-second clip is 1 minute; a 61-minute episode is 62).

Worked examples (Free/Starter prices, default memory):

Job	Calculation	Total
One-click example (26 s + 1 min 44 s)	1 + 2 = 3 min × $0.0085 + $0.002 start	$0.0275
One 60-minute podcast episode, Standard	60 × $0.0085 + $0.002	$0.512
10 interviews of 45 minutes, High accuracy	450 × $0.012 + $0.002	$5.402
1,000 voice notes of 30 seconds	1,000 × $0.0085 + $0.002	$8.502

Never charged: files that fail (404, private link, not audio, blocked site, decode error). Translation, word timestamps, subtitle files and the vocabulary hint cost nothing extra. Cap spending with Max cost per run: the Actor shortens or skips files to stay within it and tells you which ones (truncated). Apify's Free plan includes $5 of monthly credit.

Integrations (API, AI agents, Make, Zapier)

API: transcribe and get results in one call:

curl -X POST "https://api.apify.com/v2/acts/chris_elder~audio-video-transcriber/run-sync-get-dataset-items" \
  -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" \
  -d '{"sources":[{"url":"https://upload.wikimedia.org/wikipedia/commons/4/46/1941_Roosevelt_speech_pearlharbor_p1.ogg"}]}'

from apify_client import ApifyClient
client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("chris_elder/audio-video-transcriber").call(run_input={
    "sources": [{"url": "https://feeds.megaphone.fm/NATIONALAERONAUTICSANDSPACEADMINISTRATION1776343825"}],
    "maxEpisodesPerFeed": 2,
})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(row.get("episodeTitle"), row.get("wordCount"), row.get("files", {}).get("srt"))

AI agents (MCP): add https://mcp.apify.com/?tools=chris_elder/audio-video-transcriber to Claude, Cursor, VS Code or any MCP client, and your agent can transcribe any audio or video link.

Make, Zapier, n8n: use the Apify app's Run Actor action with chris_elder/audio-video-transcriber, then send the text to Google Docs, Notion, Slack or an LLM step. Schedules plus Only episodes published since give you automatic transcripts of new podcast episodes; webhooks fire when a run finishes.

Limits and good to know

• Speed (measured on Apify at the default 4 GB): Standard transcribed 20 minutes of podcast audio in about 2 minutes 20 seconds; High accuracy with word timestamps did 10 minutes in about 5 minutes 20 seconds. More memory gives more CPU and faster runs at the same price per minute.
• Not supported: YouTube, TikTok, Instagram, Facebook, X, Vimeo, Spotify and SoundCloud page links. Their terms do not allow downloading, so the Actor returns a clear (free) error. Use your own files or a direct, shared or RSS link.
• No speaker labels (diarization) yet. Open an issue if you need it.
• Defaults: 240 minutes per file, 2 GB per file, 3 episodes per feed, 100 files per run. All adjustable.
• Whisper can occasionally mishear names or produce a short repeated phrase on long silence or music. High accuracy and the vocabulary hint reduce this.
• Files are processed one at a time, in the order given.

Is it legal?

The Actor downloads only the files and feeds you give it, one request per file, and transcribes them on Apify's servers. It doesn't crawl websites, log in or bypass protection, and it refuses links to platforms whose terms forbid downloading. Make sure you have the right to transcribe the recordings (your own content, public-domain or openly licensed recordings, or content you are licensed to use). Recordings can contain personal data; you are responsible for having a lawful basis to process it. See Apify's guide: Is web scraping legal?

FAQ

Which languages are supported? Whisper detects and transcribes 99 languages. Pick one in Language to skip detection, or turn on Translate to English to get English text and subtitles from any language.

Can I transcribe a whole podcast? Yes. Paste the RSS feed or the Apple Podcasts show link and set Max episodes per feed (or a start date). Each episode is one row with its title, date and links.

How accurate is it? Standard (Whisper base) is good for clear speech such as podcasts, webinars and voice notes. Use High accuracy (Whisper small with beam search) for noisy audio, strong accents or technical vocabulary.

Is my audio sent to OpenAI or another AI service? No. The open-source Whisper model runs inside the Actor on Apify. Nothing is sent to third-party AI APIs.

How do I get subtitles for a video? Give the video file link (MP4, MOV, MKV, WebM). Each row has SRT and VTT download links you can upload to your video player or editor.

What happens if a file fails? You get a row with status: "failed" and the reason, and you are not charged for it. The run carries on with the other files.

Support, reliability and feature requests

• Failed files are free. You pay only for minutes transcribed.
• Tested on real recordings before every release (public-domain speeches, a NASA podcast feed and Apple Podcasts link, video files, broken and blocked links).
• Issues answered within 2 business days. Open an issue on the Issues tab with a sample link.
• Feature requests welcome: open an issue titled "Feature request: ...".
• Built by GPT Shop Ltd with AI-assisted development; every release is reviewed and tested before it ships.

More tools from GPT Shop Ltd

• AI Image Upscaler: batch 2x, 3x and 4x Real-ESRGAN upscaling for image links or datasets
• PDF to Markdown & Text Converter: clean Markdown and text from PDFs, with OCR for scanned pages
• Image to Text OCR: bulk text extraction from images in 100+ languages
• RSS Feed Reader: items from many RSS and Atom feeds as clean rows
• All tools: apify.com/chris_elder