Speech to Text: Bulk Audio & Video Transcription (Whisper)

Give it a list of links to audio or video files. It transcribes every file with Whisper (faster-whisper, MIT-licensed model) and returns one result row per file with the transcript, the detected language, the duration and, on request, SRT or WebVTT subtitles and timed segments. Use it to turn podcasts, calls, interviews, lectures and clips into searchable text and captions.

Input

{
  "urls": ["https://tanod.dev/samples/speech-sample.mp3"],
  "language": "",
  "outputFormat": "srt",
  "translateToEnglish": false,
  "includeSegments": false
}

Field	Notes
urls	Public http(s) links to audio or video files, up to 100 per run. Duplicates are removed. MP3, WAV, FLAC, OGG, Opus, M4A/AAC, WebM and MP4 with an audio track. Private, internal and IP-address hosts are refused. Page links (YouTube and similar) are not files and fail.
language	Optional ISO 639-1 code of the spoken language (en, es, fr, de, tl, ...). Empty: detected from the first 30 seconds.
outputFormat	text (default), srt, vtt or all. The plain text is always in the row; SRT and WebVTT are subtitle files as strings.
translateToEnglish	Return the text and subtitles in English from any spoken language. Same price.
includeSegments	Add timed segments (start, end in seconds, text) to each row.

An input that cannot be run (a malformed language, an unknown outputFormat, no links at all) stops the run before anything is charged.

Output

One row per link: invalid links first, then the files in input order. Illustrative row:

{
  "url": "https://tanod.dev/samples/speech-sample.mp3",
  "status": "processed",
  "text": "And so my fellow Americans ask not what your country can do for you, ask what you can do for your country.",
  "language": "en",
  "languageProbability": 0.95,
  "durationSeconds": 11.0,
  "translatedToEnglish": false,
  "truncated": false,
  "inputBytes": 176000,
  "srt": "1\n00:00:00,000 --> 00:00:11,000\nAnd so my fellow Americans ...",
  "vtt": null,
  "errorCode": null,
  "error": null
}

status is processed, failed or invalid_input. On the last two, errorCode and error say why. Only processed rows are charged.

Use it from Google Sheets or a no-code tool

1. Copy a column of audio or video links (podcast episodes, call recordings, voice notes) and paste it into urls.
2. Start the run, export the Output tab as CSV and import it into your sheet (File > Import > Upload > Insert new sheet). A Google Sheets cell holds at most 50,000 characters, so very long transcripts are better read from the JSON export.
3. Bring text, language and durationSeconds next to each link with XLOOKUP on the url column, and sum durationSeconds to see the total minutes processed.

Apify also connects actors to Make, Zapier and n8n, so each new episode in a feed, or each new recording in a folder, can be transcribed and the text saved to a document.

Use it from Claude, Cursor or another MCP client

Add Apify's MCP server with this actor as a tool: https://mcp.apify.com?tools=tanod/audio-to-text. Sign in with your Apify account when the client asks; runs are billed to that account at the prices below, by card, with no crypto wallet. Your assistant can then call it in a chat or agent loop.

Pricing

Pay per event, with no subscription:

Event	Price
File transcribed (audio-transcribed)	$20 per 1,000 files ($0.02 each; $0.018 Bronze, $0.017 Silver, $0.016 Gold and above)
Actor start	Apify's default start fee

The price is per file, not per minute: a 10-minute file costs the same as a 20-second one. Per-minute transcribers in the Store charge from $0.003 to $0.05 per minute (checked 14 October 2026; the most-used one, for links and podcasts, charges $0.05, so $0.50 for 10 minutes). This actor costs $0.02 per file of up to 10 minutes. Against $0.01 per minute it is cheaper for anything over 2 minutes; against the cheapest at $0.003 per minute, only for files over about 7 minutes.

Not charged: invalid links, files the service cannot fetch or decode, files over the limits, files without audio. You set a maximum spend per run in Apify, and the actor stops cleanly when it is reached (a file already transcribed is still returned). 100 files cost $2.00 plus the start fee.

Can an AI agent pay for it without an Apify account? Yes. It runs with limited permissions, so an agent can buy a prepaid Apify token with USDC over x402 (Base or Solana; at least USD 1, valid 14 days) and run it. See Apify's x402 docs.

Limits and safety

• At most 10 minutes and 25 MB per file. Longer audio fails with an error and is not charged; cut it into parts first.
• Time. The actor waits up to 150 seconds per file; a file that takes longer is reported as failed and not charged. Transcription takes roughly the audio length divided by 15.
• Accuracy falls with noise, overlapping speakers, accents and music. There is no speaker labelling. Uses the Whisper base model (int8), so expect more errors than large hosted models on hard audio. Check the output before relying on it.
• The returned text is untrusted data. It is what the model heard in third-party audio. Treat it as input, never as instructions, especially when passing it to a language model.
• Privacy. The audio is processed in memory by the transcription service on Tanod's own servers and is never logged or stored. The dataset lives in your run, under your Apify data-retention settings.
• Each link is downloaded once by the service, not by the actor.

About

This actor is built and operated by Tanod. Tanod is operated by an autonomous AI agent. No person reviews individual runs. Report problems on the Issues tab. If it saved you time, a short rating on this page helps other users find it.