Audio & Video to Text Transcriber (Whisper, No API Key)

Audio & Video to Text transcribes MP3, MP4, WAV, M4A, OGG, WEBM and podcast RSS feeds to text with OpenAI's open-source Whisper model running inside the Actor, so you need no API key. You get the full transcript, timestamped segments, and ready-made SRT and VTT subtitle files in 90+ languages, with optional translation to English. From $0.004 per audio minute.

How to use it

1. Paste direct links to audio or video files into Audio or video file URLs, or add a podcast RSS feed to Podcast RSS feeds.
2. Pick a model (Standard is fastest and cheapest, Accurate handles accents, noise and non-English speech better) and, if you know it, the language.
3. Click Start. Each file becomes one row with the transcript and links to its .txt, .srt and .vtt files.

Features

• Whisper runs in the Actor: no OpenAI key, no extra API bills, the price per minute is all-in
• Two models: Standard (Whisper base) and Accurate (Whisper small), both on the efficient faster-whisper engine with voice-activity detection
• 90+ languages with auto-detection, or force a language for short clips
• Translate to English from any supported language (same price)
• Subtitles: SRT and VTT files saved for every file, plus a plain .txt transcript
• Timestamped segments (start and end in seconds) for search, chapters and clipping
• Podcast feeds: give an RSS feed and transcribe the newest N episodes; the episode title and publish date are added to each row
• Bulk: many files per run; works with MP3, M4A, AAC, WAV, FLAC, OGG, OPUS, MP4, MOV, WEBM, MKV and most other formats
• Fair billing: failed downloads, web pages instead of files and files with no speech are not charged

Use cases

Transcribing podcasts for show notes, SEO and newsletters; meeting, interview and webinar recordings to text; subtitles for videos; feeding audio into AI, RAG and search pipelines; monitoring a podcast feed on a schedule; translating foreign-language audio to English.

Input example

{
  "mediaUrls": ["https://upload.wikimedia.org/wikipedia/commons/c/c4/GWBush_Oval_Office_Address_20010911-1-.ogg"],
  "podcastFeeds": ["https://www.nasa.gov/feeds/podcasts/houston-we-have-a-podcast"],
  "maxEpisodesPerFeed": 1,
  "model": "standard",
  "language": "auto"
}

Output example

{
  "source": "https://upload.wikimedia.org/wikipedia/commons/c/c4/GWBush_Oval_Office_Address_20010911-1-.ogg",
  "status": "ok",
  "model": "standard",
  "language": "en",
  "languageProbability": 0.992,
  "durationSec": 262.1,
  "minutesCharged": 5,
  "words": 592,
  "text": "Good evening. Today, our fellow citizens, our way of life, our very freedom, came under attack ...",
  "segments": [{"start": 0.0, "end": 8.72, "text": "Good evening. Today, our fellow citizens, our way of life, our very freedom, came under attack"}],
  "txtUrl": "https://api.apify.com/v2/key-value-stores/<storeId>/records/<name>.txt",
  "srtUrl": "https://api.apify.com/v2/key-value-stores/<storeId>/records/<name>.srt",
  "vttUrl": "https://api.apify.com/v2/key-value-stores/<storeId>/records/<name>.vtt",
  "tookSec": 8.6
}

How much does it cost to transcribe audio to text?

Pay per event: you pay only for audio that was transcribed.

• $0.004 per audio minute with the Standard model ($0.24 per hour)
• $0.008 per audio minute with the Accurate model ($0.48 per hour)

Minutes are rounded up per file. Translation to English costs the same. Files that fail to download, links that return a web page, files over your size or length limits and files with no speech are not charged.

Example: A 60-minute podcast episode costs $0.24 with Standard or $0.48 with Accurate. 100 five-minute voice notes cost $2.00 with Standard.

Apify also charges its standard run-start fee of $0.00005 per GB of run memory (a fraction of a cent per run). The exact prices are always on the Pricing tab, and you can cap any run with Maximum cost per run: the run stops when the budget is reached and the last file is transcribed only up to the minutes your budget allows (that row says "truncated": true).

FAQ

Do I need an OpenAI or other API key? No. Whisper runs inside the Actor.

How fast is it? In our tests a 60-minute podcast took about 3 minutes with the Standard model at the default 8 GB memory. The Accurate model is about 2-3 times slower.

Which model should I choose? Standard is good for clear speech such as podcasts, talks and meetings in common languages. Choose Accurate for strong accents, background noise, several speakers talking over each other, or less common languages.

Does it identify speakers? No, there is no speaker labelling (diarization). You get timestamped segments.

Can I transcribe YouTube, TikTok or Instagram links? No. This Actor only downloads direct links to media files and podcast feed enclosures. Use a link that downloads the file itself.

What are the limits? Up to 240 minutes and 1,000 MB per file by default (change Max minutes per file and Max file size).

More Tidyfold tools

Tool	Use it for
PDF OCR: Scanned PDF to Text Extractor	PDF OCR
Website Screenshot API: Bulk, Full-Page & Mobile	website screenshot API
HTML to PDF Converter API: URL & HTML to PDF	HTML to PDF
PDF Table Extractor: PDF to Excel, CSV & JSON	PDF to Excel
Invoice & Receipt Parser: PDF to JSON (No AI Key)	invoice parser
Sitemap Extractor: Get All URLs from Any Website	sitemap extractor

Use it from code or no-code tools

Every run can be started from the Apify API, the Apify Python/JS clients, or connected to n8n, Make, Zapier and Google Sheets through Apify integrations. Results download as JSON, CSV, Excel or XML.

from apify_client import ApifyClient
client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("tidyfold/audio-to-text").call(run_input={"mediaUrls": ["https://example.com/episode.mp3"], "model": "standard"})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["text"])

Notes

Only transcribe audio you have the right to use. Private and local network addresses are refused. Whisper is an open-source model (MIT licence) by OpenAI; this Actor is not affiliated with or endorsed by OpenAI. Transcripts can contain mistakes, so check anything important.

Support

Found a bug or need an option added? Open an issue on the Issues tab. Tidyfold maintains these tools daily and usually replies within 1 business day.