Pull the audio track out of any public post on X and save it on its own.
Sometimes the picture is not the point. An interview clip, a conference talk, a piece of a podcast someone posted: all of it is audio that happens to have a video attached. Saving the whole MP4 to listen to twelve minutes of speech wastes space and makes it awkward to play in the background.
This page pulls the audio stream out for you. What you get is the sound exactly as X stored it, with no re-encoding step in between, which means no extra quality loss on top of the compression the platform already applied.
How to extract Audio From Twitter/X Videos
1
Copy the post link
Open the post that contains the video, tap the share icon and choose Copy link. Threads need the link to the specific post holding the clip you want.
2
Paste it above
The tool reads what the post carries and lists the streams, including any audio-only option the platform provides.
3
Pick the audio option
Choose the entry labelled as audio only. It saves as M4A or MP3 depending on what X stored, both of which play everywhere.
M4A, MP3 and why the label matters less than you think
Audio on X is AAC, normally inside an M4A container at around 128 kbps. That is a better codec than MP3 at the same bitrate, and it plays natively on iPhone, Android, Windows, macOS, car stereos and every podcast app worth using. If a tool hands you a file called .mp3 from a source that stored AAC, it either renamed the file or re-encoded it, and re-encoding always costs a little quality.
We give you the stream as it exists. If you specifically need MP3, because some older piece of hardware demands it, convert locally with ffmpeg: ffmpeg -i input.m4a -codec:a libmp3lame -q:a 2 output.mp3 produces a good variable bitrate file in a few seconds.
What the audio will sound like
X targets roughly 128 kbps AAC for the audio in most posts. For speech that is genuinely transparent, you will not hear a difference from the source. For music it is adequate rather than good, and on decent headphones you may notice cymbals and reverb tails losing definition.
None of that can be recovered afterwards. Bitrate is a ceiling set when the uploader posted. Converting to a higher bitrate MP3 makes a bigger file containing exactly the same information, which is a common misunderstanding worth being clear about.
Fair use and music
Saving a talk, a lecture clip or a piece of commentary to listen to on a commute is ordinary personal use. Extracting a song from a post and treating it as a music library is a different activity, and one where the rights holder has a real complaint.
Music posted to X is usually there under a licence that covers playback on X and nothing else. Pulling the audio out and republishing it elsewhere is what triggers automated copyright claims on other platforms. Our disclaimer covers where personal use ends in more detail.
Twitter to MP3 questions
Normally AAC at around 128 kbps, which is what X stores. That is transparent for speech and reasonable for music. The bitrate was fixed when the video was uploaded, so no tool can improve it after the fact, and converting to a higher bitrate only makes the file larger.
Usually M4A, because the audio inside an X video is AAC and M4A is its natural container. It plays on every phone, computer and car stereo made in the last fifteen years. Convert it locally if a specific device insists on the .mp3 extension.
No, because those posts have no audio stream at all. X stores GIF-style clips as silent looping MP4s. If the post plays silently in the app, there is nothing to extract, and our Twitter GIF page explains why that format works the way it does.
Recorded Spaces use a different system from video posts and are not supported here. If a Space recording has been posted back to a timeline as a video clip, that clip works like any other post and its audio can be extracted normally.
For personal listening, generally yes, in the same way that saving a public video to watch offline is. Redistributing the audio, especially music, is where copyright bites. The creator owns the recording regardless of which part of it you saved.
X applies loudness normalisation during playback, not to the stored file. The raw stream is what the uploader supplied, which may be quieter or louder. Any audio editor can normalise it in a few seconds if you are collecting clips that need to match.
Yes. Length is not a limit here, though very long posts take longer to read because there is more to process. Audio-only files stay small regardless: an hour of 128 kbps AAC is around 55 MB, against several gigabytes for the same thing with video attached.
Not if you pick the audio-only option. That entry contains a single audio stream with no video track, which is why it is a fraction of the size. Choose one of the resolution options instead if you want the picture too.
