How to Extract Audio from a Zoom Recording (MP4 → MP3 in 60 Seconds)

You finished a Zoom meeting, downloaded the recording, and now you want just the audio — to drop into Otter for transcription, edit into a podcast clip, attach to an email under the 25 MB cap, or feed to a Whisper model. The MP4 is sitting there at 400 MB; the audio is only 30 MB of that, and there is no obvious button in Zoom to download audio-only after the fact.

This is one of those tasks where the right answer depends on whether you have ffmpeg, whether the file is sensitive, and how often you'll do it. Below are the three methods that actually work, ranked by speed.

Method 1: Zoom's Hidden "Audio Only" Download

If you have not yet downloaded the recording — and the meeting was recorded to the cloud (not local) — Zoom already split the audio into a separate file. Most users miss this because the UI buries it.

1. Sign into the Zoom web portal: zoom.us/recording.
2. Click the recording's title (not the "Share" button).
3. You'll see a list of files: usually Recording.mp4, audio_only.m4a, chat.txt, and a transcript if you enabled it.
4. Click audio_only.m4a to download just the audio.

The audio_only.m4a file is AAC-encoded mono, roughly 8–12 MB per hour of meeting. Drop it directly into Otter, Whisper, or any podcast editor — most modern tools handle M4A natively.

When this fails: locally-recorded meetings (the local Zoom client saves only zoom_0.mp4, no separate audio file), or recordings older than your account's retention window (default 120 days for paid plans, 30 days for free). For both, fall back to method 2 or 3.

Method 2: ffmpeg One-Liner (Fastest If You Have It)

If you have a Mac or Linux machine with ffmpeg installed, audio extraction is one command and runs in ~5 seconds for a 1-hour meeting:

ffmpeg -i zoom_recording.mp4 -vn -acodec copy output.m4a

Breakdown:

• -i zoom_recording.mp4 — input file.
• -vn — no video (skip video stream).
• -acodec copy — copy the audio stream without re-encoding (lossless, instant).
• output.m4a — output file (keep the extension matching the codec; Zoom uses AAC).

If you'd rather have MP3 (broader compatibility, especially for older transcription services or email clients), re-encode:

ffmpeg -i zoom_recording.mp4 -vn -acodec libmp3lame -b:a 128k output.mp3

128 kbps mono is plenty for voice; 64 kbps works if you're storage-constrained. Higher bitrates than 128 are wasted on Zoom audio (the source is itself ~64 kbps AAC).

Install ffmpeg if you don't have it:

• macOS: brew install ffmpeg
• Ubuntu/Debian: sudo apt install ffmpeg
• Windows: winget install Gyan.FFmpeg or download from ffmpeg.org

Why ffmpeg is the gold standard:

• Local, no upload — your meeting audio never leaves your machine. Critical for HIPAA / financial / legal recordings.
• Lossless — -acodec copy doesn't re-encode, so quality is identical to the source.
• Batchable — drop it in a shell loop to process 100 recordings overnight.

for f in *.mp4; do
  ffmpeg -i "$f" -vn -acodec copy "${f%.mp4}.m4a"
done

Method 3: Browser-Based Extractor (No Install)

If you don't have ffmpeg and don't want to install it for a one-off, browser-based tools do the job. The trustworthy ones run entirely client-side via ffmpeg.wasm — no upload, no server processing, no file size cap beyond your browser's memory.

orangebot.ai/tools/audio-extractor is the one we built and run; it runs ffmpeg.wasm in the browser, supports MP4 / MOV / WebM / MKV inputs, and outputs MP3 / WAV / M4A. The file never leaves your tab — you can verify by opening the browser's Network panel during the export and watching for zero uploads.

Workflow:

1. Open the tool, drag the Zoom MP4 onto the drop zone.
2. Pick output format (MP3 for compatibility, M4A for size, WAV for downstream editing).
3. Wait 5–20 seconds depending on input size and your CPU.
4. Click download.

Comparison vs ffmpeg CLI: slower (browser is roughly 3–5× slower than native ffmpeg), but zero install and zero command-line knowledge. For one or two files a month, the trade-off is fine. For dozens, install ffmpeg.

Other options in the same category:

• online-audio-converter.com — popular, but uploads your file to their servers. Don't use for sensitive recordings.
• convertio.co — same upload-based model, with file-size limits on the free tier.
• CloudConvert — paid, server-side, with stronger SLAs if you're processing for clients.

The pattern: free server-side converters are convenient but bad for anything resembling a private meeting. Client-side WASM converters give you both privacy and zero install — pick those.

Privacy: The Decision Tree

Zoom recordings often contain sensitive information you don't want sitting on someone else's server:

• Customer calls — recorded consent or not, processing on a third-party server can break your privacy policy commitments.
• Internal strategy meetings — IP, headcount discussions, M&A signals — pure leak risk.
• Healthcare / finance / legal — HIPAA, GLBA, attorney-client privilege all imply audio data should not transit unaffiliated cloud services without a BAA / contract.

The right call by sensitivity:

If you're regularly transcribing internal calls, the high-leverage workflow is: ffmpeg-extract audio locally → run Whisper locally with a quantized model → keep the transcript on disk. Total cost: zero dollars, zero data egress.

Bonus: Audio Quality Tips

Zoom's source audio is heavily compressed (32–64 kbps AAC mono, often with noise suppression baked in). A few things to know:

• Re-encoding loses quality. If you re-encode AAC to MP3 you lose another ~3% intelligibility. Use -acodec copy whenever possible.
• 128 kbps is the realistic ceiling for any output. Encoding the source at 320 kbps mp3 doesn't make it sound better — it just makes the file bigger.
• Stereo is wasted on Zoom voice. Most Zoom recordings are mono; don't pay for stereo.
• Transcription accuracy is bottlenecked by source quality. A clean 64 kbps mp3 from a headset mic will transcribe better than a 320 kbps mp3 of laptop-mic audio.

If you regularly process Zoom audio for transcription, the single biggest quality win is asking participants to wear headsets — not anything you can do in post.

FAQ

Will extracting audio break Zoom's terms of service? No. The recording is yours (or your account's). Extracting audio is a format conversion, not a TOS issue. Sharing the recording is governed by the consent rules of the jurisdictions of the people in it — that's a separate question.

Why is audio_only.m4a mono? Zoom downmixes all participant audio into a single mono track to save storage. You can't recover stereo from it — even with separate audio tracks (a feature for paid accounts, off by default), it's per-participant mono, not stereo.

What about Zoom Phone or Zoom IQ transcripts? If Zoom IQ is enabled, your meeting already has a server-side transcript at zoom.us/recording. Download the .vtt or .txt directly — no audio extraction needed for that workflow.

Can I extract audio from a Teams or Google Meet recording the same way? Yes. Teams recordings are MP4 stored in OneDrive (ffmpeg -i meeting.mp4 -vn -acodec copy out.m4a), Google Meet recordings are MP4 in Drive (same command). The browser tools also handle both.

What's the smallest possible MP3 that's still understandable? ~32 kbps mono. Below that, sibilants get crunchy. For voice transcription you can go as low as 16 kbps with Opus codec, but compatibility is limited.

If you do this often, install ffmpeg today and stop fighting the GUI. For a one-off, the browser-based audio extractor is two clicks and respects your privacy. And if you're working with other media formats, the media converter handles MP4/MOV/WebM/MKV/AVI conversions with the same client-side approach.