Video to MP3
Keep the sound, drop the picture. Nothing is uploaded, whatever the size.
No size limit and no queue: the file is read from your own disk.
- 100% freeNo account, no daily quota, no paid tier.
- Nothing is uploadedFiles are processed on your device and never reach a server.
- No catch on the outputNo watermark, no forced downscaling, no size cap.
- Updated in the openEvery fix and every new tool is written down on the changelog page.
What this tool does
Drop a video in and this page gives you the audio from it. Two ways out. MP3 is the default, because that is the file most people came here for: the audio is decoded and re-encoded with LAME at the bitrate you pick. The second way is to take the track out untouched. The sound in a video is already a separate, finished audio stream, usually AAC, so it can be lifted straight into an audio container without a single sample being decoded. That path is lossless and finishes in about the time it takes to read the file.
Nothing is uploaded. That matters here more than it would for a text tool: a half-hour recording is a gigabyte, and a converter that works by uploading has to put a ceiling somewhere, usually a few hundred megabytes on the free tier with the larger files behind a subscription. There is no ceiling here because there is no upload. The file is read from your own disk and the work happens in the tab, so a two-hour recording is not a different case from a two-minute one.
How to use it
Drag a file onto the page, or click to pick one. Before anything is converted, the page reads the container and tells you what is in it: codec, channels, sample rate and length. That answers the question that sinks most attempts at this, which is whether the video has usable audio at all. Screen recordings made with the microphone off, and clips stripped by a messaging app, arrive with no audio track, and you find that out here in a second rather than after a long conversion.
Pick MP3 or the original track, press the button, and listen to the result in the player before you download it. The bitrate control sits under Settings and offers 128, 192 and 320 kbps. The default is 192 rather than 128 because the audio in a video is usually AAC at 128 to 192 kbps, and MP3 is the older format: encoding to MP3 at the same number as the source loses a little you can hear on music.
If the video carries more than one audio track, which happens with dubbed films and commentary tracks, a selector appears and lists them with their language codes. With a single track there is no selector, because a menu with one item in it is just something to click past.
Things worth knowing
An MP3 comes out slightly longer than the source. An MP3 frame holds exactly 1152 samples, so the final frame is padded, and the encoder adds a short delay of its own; three seconds of input measured 3.048 seconds out. That is the format, not a fault, but it matters if you are lining the audio up against the original video frame by frame. The untouched path has no such drift, since nothing is re-timed. Browsers also cannot make an MP3 on their own: the WebCodecs audio encoder handles Opus, AAC and FLAC but has no MP3 mode, so this page carries its own LAME encoder, and choosing MP3 downloads a little more code than the lossless path does. That path works in places the MP3 path does not, because copying a track needs no decoder for it. The audio bytes come out identical to those inside the video, so a checksum taken from both matches.
Background: where MP3 came from
MP3 is MPEG-1 Audio Layer III, standardised in ISO/IEC 11172-3 in 1993, and developed largely at Fraunhofer IIS under Karlheinz Brandenburg. It won the format war of the early internet on one property: it was the first codec good enough at low bitrates to make music worth downloading over a modem. It is no longer the best codec at any bitrate, and AAC or Opus beat it comfortably, but it plays on everything ever built, which is why it stays the default here. The patent pool that made it awkward for open-source software for two decades ran out in 2017, when Fraunhofer ended its licensing program, which is part of why a browser page can now encode one without asking anybody.
Where people use it
The everyday case is a lecture, interview or podcast recorded as video where only the talking matters, and an audio file is a twentieth of the size and playable on anything. Musicians pull the sound off a rehearsal video. People preparing something for transcription convert first, because most transcription tools take audio and choke on a large video. Language learners keep the audio from a clip to listen to on a walk. If the video itself is what needs fixing rather than the sound, rotate video turns a sideways clip upright without re-encoding it. If the sound is what you want gone rather than kept, remove audio from video leaves the picture untouched.
Frequently asked questions
Is the audio quality reduced?
On the MP3 setting, yes, slightly: MP3 is a lossy format and the audio is encoded again. Choose the original track instead and nothing is re-encoded at all, so the result is identical to the audio inside the video.
Is there a file size limit?
No. The file is read from your own disk and never uploaded, so a two-hour recording works the same as a two-minute one. It just takes longer on the MP3 setting, because every second of audio has to be encoded.
Why is the original track an M4A and not an MP3?
Because that is what is already inside the video. Most videos carry AAC audio, which belongs in an M4A container. Turning it into an MP3 would mean re-encoding it, which is exactly what that option exists to avoid.
The page says my video has no audio track.
Then the file genuinely has none. Screen recordings made with the microphone off, and clips that have been through an app that strips audio, both arrive this way. There is nothing to extract.
Which bitrate should I pick?
192 kbps for anything with music in it, which is the default. 128 is fine for speech and makes the file a third smaller. 320 only pays off if the audio in the video was high quality to begin with.