Audio Extractor — Video to MP3, WAV, FLAC (Speech-to-Text Ready)

Pull the audio track out of videos in bulk and save it as MP3, M4A (AAC), Opus, WAV or FLAC. One switch gives you the format speech recognition engines like Whisper work best with (WAV, mono, 16 kHz), and optional loudness normalization fixes recordings that are too quiet or too loud. Metadata is removed from every file.

Real output of this Actor: a quiet street recording raised from -36 to -17 LUFS, and measured file sizes per format. Audio: Qviri, CC BY-SA 4.0, Wikimedia Commons.

Why use it

• 🗣️ Better transcripts, lower cost: speech-to-text APIs charge by upload size or duration and work best with 16 kHz mono audio. Sending a 1 GB video when 110 MB of WAV (or 25 MB of Opus) carries the same speech wastes bandwidth and time.
• 🔊 Even volume: loudness normalization to the podcast standard (-16 LUFS) makes quiet phone recordings and loud webinars sound the same.
• 🎧 Every common format: MP3 for compatibility, M4A for Apple and podcasts, Opus for the smallest files, FLAC/WAV for lossless archives and editing.
• 📦 Bulk and automated: hundreds of files per run, from URLs, Google Drive or Dropbox links, on a schedule or from a webhook.
• 🔒 Metadata removed: no device, software or location tags in the output.
• 🧩 No FFmpeg setup: send URLs, get audio links back.

Who uses it

Sector	Typical use
AI & speech-to-text pipelines	Prepare audio for Whisper, Deepgram, AssemblyAI, Google or Azure Speech
Podcasters & content creators	Turn video episodes, lives and interviews into podcast audio
Education & e-learning	Audio versions of lectures and courses for listening on the go
Companies & legal	Archive the audio of meetings, hearings, depositions and calls
Media monitoring & research	Extract speech from news, social and broadcast videos for analysis
Call centers & QA	Normalize recorded video calls before transcription and scoring
Data teams	Build audio datasets (speech, sound events) from video collections

Formats

Measured on real output (size per hour of audio):

Format	Setting	Size / hour	Best for
Opus	64 kbps	~25 MB	Smallest files, excellent for speech
MP3 (default)	128 kbps	~55 MB	Plays on any device
M4A (AAC)	128 kbps	~56 MB	Apple devices, podcast platforms
Speech-ready WAV	16 kHz, mono	~110 MB	Speech-to-text engines (Whisper & co.)
FLAC	lossless	~540 MB	Archive without quality loss
WAV	original rate	~660 MB (48 kHz stereo)	Audio editing software

Lossy formats use the Bitrate you choose (32–320 kbps).

Input

Field	Type	Default	Description
Media URLs (mediaUrls)	array	—	Links to videos or audio files. Public Google Drive and Dropbox share links are converted automatically. Up to 4 GB per file.
Speech-to-text ready (speechReady)	boolean	false	Overrides format, rate and channels with WAV, 16 kHz, mono
Format (format)	string	mp3	mp3, m4a, opus, wav or flac
Bitrate (kbps) (bitrateKbps)	integer	128	For MP3, M4A and Opus (32–320)
Sample rate (sampleRate)	string	original	original, 16000, 22050, 44100 or 48000
Channels (channels)	string	original	original, mono or stereo
Normalize loudness (normalizeLoudness)	boolean	false	EBU R128 normalization to -16 LUFS
Max minutes per file (maxDurationMinutes)	integer	180	Only the first N minutes are extracted and billed

Example:

{
  "mediaUrls": [
    "https://example.com/webinar.mp4",
    "https://www.dropbox.com/s/abc123/interview.mov?dl=0"
  ],
  "format": "mp3",
  "bitrateKbps": 96,
  "channels": "mono",
  "normalizeLoudness": true
}

Supported inputs: MP4, MOV, WebM, MKV, AVI, FLV, MPEG-TS, and audio files such as M4A, OGG, WAV, FLAC and MP3. The first audio track of each file is used.

Output

One dataset row per file. The audio file is stored in the run's key-value store and linked in audioUrl.

{
  "sourceUrl": "https://example.com/webinar.mp4",
  "audioUrl": "https://api.apify.com/v2/key-value-stores/.../records/audio-0001.mp3",
  "format": "mp3",
  "durationSeconds": 3605.2,
  "sizeMB": 55.1,
  "sourceAudioCodec": "aac",
  "truncated": false,
  "billedMinutes": 61
}

Field	Meaning
audioUrl	Download link of the audio file
format, durationSeconds, sizeMB	What you got
sourceAudioCodec	Audio codec of the input (aac, opus, mp3, pcm…)
truncated	true when only the first Max minutes per file were extracted
billedMinutes	Started minutes billed for this file
error	Present only when a file failed. Files without an audio track return a clear error and are not billed

Pricing

Pay per started minute of audio extracted, plus a tiny per-run start fee. No subscription. Failed files are free. Apify plan discounts apply automatically (see the Pricing tab).

Example	Minutes billed
30-second clip	1
1-hour webinar	60
100 short videos of 2 minutes	200

Use Max minutes per file or the run's maximum cost setting to cap spending; the Actor stops cleanly when the limit is reached.

Use it from code or an AI agent

Python: extract, then transcribe

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("adorable_partial/audio-extractor").call(
    run_input={"mediaUrls": ["https://example.com/webinar.mp4"], "speechReady": True}
)
audio_urls = [i["audioUrl"] for i in client.dataset(run["defaultDatasetId"]).iterate_items() if "audioUrl" in i]

# Optional: send the audio to the Whisper transcriber Actor
t = client.actor("adorable_partial/whisper-audio-video-transcriber").call(run_input={"mediaUrls": audio_urls})
for item in client.dataset(t["defaultDatasetId"]).iterate_items():
    print(item.get("text", "")[:200])

JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('adorable_partial/audio-extractor').call({
    mediaUrls: ['https://example.com/episode-12.mp4'],
    format: 'mp3',
    normalizeLoudness: true,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items[0].audioUrl);

HTTP

curl -X POST "https://api.apify.com/v2/acts/adorable_partial~audio-extractor/run-sync-get-dataset-items?token=<YOUR_APIFY_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"mediaUrls": ["https://example.com/webinar.mp4"], "format": "opus", "bitrateKbps": 48}'

No-code and agents: use the Apify modules in Make, Zapier or n8n, trigger runs from webhooks or schedules, or let an AI agent call it through the Apify MCP server.

FAQ

Which format should I pick for transcription? Turn on Speech-to-text ready (WAV, 16 kHz, mono). If your transcription service limits upload size, use Opus 32–64 kbps mono instead: about 4–6× smaller with practically the same accuracy for speech.

Does normalization change the content? No. It only adjusts the volume so the whole file sits around -16 LUFS without clipping peaks.

Can it extract from YouTube or social media links? It needs a direct link to the file (or a public Google Drive / Dropbox link). Page URLs of streaming sites are not supported.

What about videos with several audio tracks? The first audio track is extracted.

How long can files be? Up to 4 GB per file; up to Max minutes per file (default 180) are extracted.

Is my data kept? Files are processed inside your run and stored only in your own Apify storage, under your account's data-retention settings.

Related Actors

• Audio & Video to Text (Whisper) — transcripts, SRT and VTT subtitles
• Video Optimizer — reduce FPS, resize and compress videos
• Video Frame Extractor for AI — turn videos into clean image datasets
• Image & Video Anonymizer — blur faces and license plates (GDPR / LGPD)

Support

Need another format or setting? Open an issue in the Issues tab.