How to Extract Audio from a Video (MP4, MOV, MKV) Without Uploading It

Every video file is two recordings sharing one container: a picture track and a sound track. Most of the time the picture is the heavy one, and quite often it is the one you do not need. Extracting the audio is simple once you know what is inside the file, and it can be done entirely on your own device.

Why pull the audio out at all

The reasons are mundane and extremely common:

• Podcast and interview clips. Meetings and interviews get recorded as video because that is what the app offered. The audio is the content; the video is dead weight.
• Lecture recordings. A ninety-minute lecture at 1080p can run to 2 GB. The same lecture as a 96 kbps MP3 is about 65 MB and plays on a commute without draining a phone.
• Ringtones and alert sounds. A line from a film clip, a child laughing, a ten-second riff.
• Music from a concert or rehearsal clip that you want in a playlist rather than a photo library.
• Transcription. Many transcription tools accept audio only, or charge by the megabyte.

What is actually inside the video

Almost every MP4 and MOV from a phone, screen recorder, camera or video site carries AAC audio. It is the standard partner of H.264 and HEVC video, and it has been for twenty years. MKV is more varied: AAC and Opus are common, but rips from discs often carry Dolby AC-3 or DTS. WebM uses Opus or Vorbis.

That matters because there are two ways to “extract” a track:

1. Copy it unchanged into an audio-only container. The AAC stays exactly as it was, and you get an .m4a file. This is lossless but needs a command-line tool.
2. Decode it and encode it as MP3. This is what most converters do, including Tinyvert, because MP3 is the format everything plays.

What converting AAC to MP3 does to the sound

AAC and MP3 are both lossy. Decoding AAC back to raw samples is exact, but encoding those samples as MP3 discards detail a second time. This is lossy-to-lossy conversion, and the honest summary is: the MP3 can never sound better than the AAC it came from, and at a generous bitrate it will not sound noticeably worse.

The real danger is stacking a low bitrate on a low bitrate. A 96 kbps AAC voice note pushed through a 64 kbps MP3 will sound swirly. The same source at 128 kbps or higher is fine. When in doubt, match the source or go one step up.

Choosing a bitrate

The Bitrate menu on Tinyvert’s audio tools runs from 64 to 320 kbps. Rough guidance, with the size you get per minute of audio:

Two caveats. If the source track is 128 kbps AAC, choosing 320 kbps buys you nothing but a bigger file. And if you plan to edit the audio afterwards in Audacity, GarageBand or a DAW, skip MP3 for now: pick WAV or FLAC in the Audio Converter so you take one lossy hit at the end rather than two, then export the final MP3 from the editor.

Doing it in your browser with Tinyvert

1. Open MP4 to MP3 for .mp4 and .m4v files, or Video to MP3 for .mov, .mkv, .webm and .ts as well.
2. Drop up to five files. Each one gets its own row and progress bar.
3. Choose a bitrate from the Bitrate menu. The default is 192 kbps.
4. Convert. Your browser’s own decoder reads the audio track and the picture is discarded. The MP3 is written by the LAME encoder compiled to WebAssembly, about 300 KB, fetched once the first time you need it. Sample rate and channel count are preserved, and any title, artist or cover-art tags are carried across.
5. The row reports the size saving once it finishes, and with the video gone it is usually 90 percent or more. Download each MP3 individually, grab the lot as a ZIP, or use the Share button on a phone to send it straight to another app.

If you want WAV, FLAC, OGG or M4A instead of MP3, the Audio Converter accepts the same video inputs and adds an Output format menu.

The audio and video tools rely on the WebCodecs API, so they work in Chrome, Edge, Brave and Opera on desktop and Android, Safari 17 or newer, and Firefox 130 or newer. Decoding AAC depends on codecs your browser or operating system ships, which the mainstream ones do. Tracks the browser cannot decode, AC-3 and DTS in particular, will not convert; those are a job for ffmpeg below. Large files are limited by your device’s memory, since the input and output both sit in RAM, so a multi-hour 4K recording is better handled on a laptop than a phone.

A note on ringtones

Android accepts an MP3 directly: drop it in the Ringtones folder or pick it from the sound settings. iPhones want an .m4r, which is nothing more than an AAC file in an MP4 container with a different extension, kept under 40 seconds. Convert to M4A with the Audio Converter, rename the extension, and add it through Finder or iTunes. If the clip needs cutting first, run the video through Trim Video before extracting; the cut snaps to the nearest keyframe before your start time, so give yourself a second or two of slack.

Alternatives on the desktop

VLC can do this without installing anything else. Go to Media, then Convert / Save, add the file, pick the “Audio - MP3” profile and a destination, and press Start. The built-in profile is 128 kbps; edit it if you want more. It works, but it is slow to set up and easy to get wrong on the first attempt.

ffmpeg is the one-liner everyone eventually learns:

ffmpeg -i in.mp4 -vn -c:a libmp3lame -b:a 192k out.mp3

-vn drops the video, -c:a libmp3lame selects the LAME encoder (the same one Tinyvert runs in WebAssembly), and -b:a 192k sets the bitrate. To keep the AAC exactly as it is, which Tinyvert cannot do, swap the encoder for a copy:

ffmpeg -i in.mp4 -vn -c:a copy out.m4a

That version finishes in a fraction of a second because nothing is decoded, and it also handles AC-3 and DTS tracks if you change the extension to .ac3 or .dts.

A note on privacy

Video recordings are personal in a way that spreadsheets are not. Meetings, lectures with students’ names in them, family footage, a doctor’s explanation you recorded to listen to again. Most web-based converters send the entire file to a server, where it waits in a queue and lands in someone’s logs.

Tinyvert reads the file into your browser’s memory and converts it with code running on your own device. Nothing leaves the machine, nothing is logged, nothing is stored. After the first visit a service worker caches the page and the converter, so it keeps working with no connection at all. The free tier handles five files per batch with no sign-up and no watermark; Pro is a one-time US$9 upgrade that removes the batch limit.

Pick the bitrate for the content, keep the original video, and the rest takes a few seconds for a typical file.