Bulk YouTube & Video Transcriber with SRT Subtitles

Transcribe a pile of videos without collecting links. Search YouTube by topic, point it at a whole channel or playlist, or paste links from YouTube, TikTok and Instagram. MP3, WAV and MP4 files work too. Every item comes back as timestamped text plus SRT and VTT subtitle files.

Quick start

Press Start and the sample transcribes a short public audio clip so you can see the output. Then type a topic into Search YouTube (say, cold email tips), set Videos per search to 25, and run again.

What you get

• Full transcript text, detected language, duration and word count per item
• Timed segments, plus SRT and VTT subtitle files saved for each item
• Source type and URL, and a clear status and error per item, so failures are easy to rerun
• PDF and image links turned into text, page by page, with OCR for scans (beta)
• A run summary with totals, successes and failures

Good for

• Transcribing a creator's back catalogue to search or quote it
• Subtitles for a batch of clips
• Research on what people say about a topic across dozens of videos
• Turning podcast episodes into text for show notes

Pricing

• $0.0025 per file transcribed, any length, until 29 October 2026
• From 29 October 2026: $0.005 per file plus $0.003 per started audio minute
• Failed downloads and skipped files are not billed

You pay only for results that come back. Set a max cost per run in the run options and the run stops cleanly when it hits it. Normal Apify platform usage applies.

Input

Field	What it does
youtubeSearch	Type a topic, like podcast marketing tips or Lex Fridman. We find the top videos and transcribe them all. No links to copy.
videosPerSearch	How many results to take for each search term or channel.
channelUrls	Paste one channel or playlist link (youtube.com/@mkbhd, a TikTok profile) and we transcribe its latest videos, up to the number above.
urls	YouTube, TikTok, Instagram or any page with a video. Each file costs $0.005 plus $0.003 per started audio minute.
mediaUrls	Direct links to .mp3, .wav, .mp4 and similar files. The sample is a 11-second clip so you can see the output.
uploadedFiles	Links to files you uploaded to Apify storage or any file host.
documentUrls	Public PDF or image links to turn into text, page by page, with Markdown. Pages without a text layer get OCR. Links ending in .pdf, .png or .jpg in other fields are sent here automatically.
language	Two-letter code such as en, es or hi. Leave blank and we detect it.
prompt	Spellings the transcript should get right: product names, people, acronyms.
includeSubtitles	Saves timed subtitle files for every finished item.
concurrency	How many items run in parallel. Lower it if a site starts blocking downloads.
maxItems	Upper limit on links accepted in one run, including search results.
maxFileSizeMb	Larger direct files are skipped before transcription and not billed.
ocr	OCR reads text from scanned pages.
ocrLanguage	Tesseract language code. eng is installed.
maxPages	Longer files are skipped with an error row and not billed.
transcription	Leave as is. Only for teams who want to bill a Whisper-compatible endpoint of their own: model, baseUrl, apiKey.

Limits

• Public media only. Private, members-only and login-gated videos fail with an error row.
• Instagram serves some videos inconsistently to logged-out visitors.
• Direct files over 24 MB are skipped by default; raise Max file size if you need to.

Run it from code

Same input as the Console sample:

{
  "urls": [],
  "mediaUrls": [
    "https://raw.githubusercontent.com/ggerganov/whisper.cpp/927cfce34f31707e17f2bff35c349632fb9e2c3a/samples/jfk.wav"
  ],
  "transcription": {
    "model": "whisper-1"
  }
}

curl -X POST "https://api.apify.com/v2/acts/physealabs~bulk-media-transcriber/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" -d @input.json

from apify_client import ApifyClient
items = ApifyClient("YOUR_APIFY_TOKEN").actor("physealabs/bulk-media-transcriber").call(run_input={"urls": [], "mediaUrls": ["https://raw.githubusercontent.com/ggerganov/whisper.cpp/927cfce34f31707e17f2bff35c349632fb9e2c3a/samples/jfk.wav"], "transcription": {"model": "whisper-1"}})

It also works from Zapier, Make, n8n and Apify schedules.