DEV Community

Nikita Iakovlev
Nikita Iakovlev

Posted on

Audio and video to text via API: transcribe a 45-minute podcast for under $0.50

#speechtotext #ai #api #podcast

Most speech-to-text APIs make you do the plumbing yourself: download the media, strip the audio from video, cut long files under the size limit, send parts, stitch timings back together. If you just have a list of links — podcast episodes, meeting recordings, lecture videos — that's a lot of glue code for one text file.

One link in, one transcript out

Speech to Text on Apify takes MP3, M4A, WAV, FLAC, OGG, MP4, MOV or WEBM — any link or uploaded file with sound — and does the whole pipeline: download, audio extraction, splitting, recognition with Whisper large v3, and re-stitching timings. A three-hour recording is one input and one row of output.

Each row contains:

  • the clean transcript with timed passages
  • ready SRT or WebVTT subtitles
  • retrieval chunks that keep the moment each passage was spoken
  • language detected, 90+ languages supported
  • optional speaker labels 
curl -X POST "https://api.apify.com/v2/acts/lergassy~speech-to-text/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"urls":["https://example.com/episode-42.mp3"]}'
Enter fullscreen mode Exit fullscreen mode

Cost (from 4 October 2026)

  • $0.009 per started minute with the built-in recognition key — nothing to sign up for
  • $0.004 per minute if you bring your own Groq key (you then pay Groq directly, a fraction of a cent per minute)
  • $0.002 per file processed
  • speaker labels: +$0.006 per minute, only when switched on

A 45-minute episode is about $0.41 with the built-in key (45 × $0.009 + $0.002), or about $0.18 with your own key.

Video pages (YouTube, TikTok, Instagram, X, Vimeo) whose audio has to be pulled through a residential exit node carry an extra $0.09 per link — direct file links never pay it. For YouTube videos that already have captions, the transcript scraper from my earlier post on YouTube transcripts is far cheaper.

When to use it

Podcast show notes and search, meeting and interview archives, turning a course library into a searchable knowledge base, subtitles for your own videos.

Limits

  • It transcribes what is audible: heavy music or crosstalk lowers accuracy, as with any Whisper-based tool.
  • Private or login-only media can't be downloaded — use a direct file link or upload the file.

Disclosure: links to Apify in this post use my referral code.

Top comments (0)