You can convert a video to MP3 without uploading it by running the encoder inside the browser tab: the file is read from your disk into page memory, the audio track is extracted there, and the result is handed back as a download. Nothing is transmitted to a server.
The tool takes MP4, WebM, MOV, MKV and AVI input and writes MP3, M4A or WAV. MP3 and M4A let you pick 128, 192 or 320 kbps, with 192 kbps as the default; WAV is uncompressed 16-bit PCM and ignores the bitrate setting, which makes it roughly ten times larger than the same audio as MP3. The video stream is dropped entirely, so a 200 MB clip typically becomes a few megabytes of audio. The first run downloads the FFmpeg engine, about 30 MB of code, and after that the tool works offline.
Upload-based converters are everywhere, and they all ask for the same thing before they help: the file itself. For a downloaded music video that is a mild annoyance. For a recorded meeting, a lecture with participants' names on screen, or an interview about something confidential, handing over the whole file just to keep the audio is a poor trade. This guide covers what a local extraction actually does, which settings change the result, and where the tool stops.
Video to MP3
Drop in a clip, choose MP3, M4A or WAV, set the bitrate, and download the audio. The encoder runs in the page, so the video is never uploaded.
Why the upload step is the part worth removing
The conversion itself is not the risk. An audio track is a small, harmless thing. The risk is that the file you would have to hand over also contains everything else that was on screen: a customer name in a call recording, a slide marked internal, a whiteboard photo of a plan. Extracting audio locally means the extra content never travels, because the extraction happens before anything could be sent.
The practical difference between the two approaches is not speed or quality. It is whether you can name every place your file exists afterwards.
| Aihangsoft | Upload-based converter | |
|---|---|---|
| Where the video goes | Stays in the browser tab | Sent to their servers |
| Size and daily limits | None beyond your own memory | Often capped, sometimes watermarked |
| Works offline | Yes, after the first load | No |
| Account | Not needed | Sometimes required above a size threshold |
| What you can verify | The network panel stays quiet | Nothing about retention or deletion |
What goes in and what comes out
Input is any file the browser recognizes as a video, and the tool names the five common containers up front so you know what is expected: MP4, WebM, MOV, MKV and AVI. Output is deliberately narrow, because three well understood formats cover nearly every case.
| Output | Codec | Bitrate control | Best for |
|---|---|---|---|
| MP3 | libmp3lame | 128, 192 or 320 kbps | Playback anywhere, voice notes, podcasts |
| M4A | AAC in an MP4 container | 128, 192 or 320 kbps | Smaller files at the same quality |
| WAV | 16-bit PCM | Not applicable | Editing, mixing, re-encoding later |
Choosing WAV hides the bitrate control, because uncompressed audio has no bitrate to set. The tool states that in the panel rather than leaving you to wonder why the option disappeared. Everything else about the output is fixed: the audio is re-encoded in one pass, the video stream is discarded rather than hidden, and no metadata from the container is carried over.
How to extract audio from a video without uploading it
- Add the video. Drop an MP4, WebM, MOV, MKV or AVI file into the drop zone, or click to browse. The clip is read on your device, and the tool prints its size and duration once the metadata loads.
- Pick the format and bitrate. MP3 at 192 kbps is the default and a safe choice. Switch to WAV if the audio will be edited, or drop to 128 kbps for speech you only need to hear.
- Extract and download. Press Extract audio, watch the progress bar, listen to the preview in the result panel, then download. The panel shows the source size, source duration, audio duration and output size so the trade is measurable.
The result panel keeps the source video in memory only for the session, and the original file on disk is never modified. Reset clears the panel and revokes the output URL, which is a useful habit if you extracted something you would rather not leave a preview of on screen.
Reading the two duration fields
The panel reports the source duration and the audio duration separately. They normally match. When the source is a file the browser cannot parse metadata for, both can show a dash, and the conversion still proceeds: the encoder reads the stream itself rather than trusting the container header. A dash is a display limitation, not a sign that the file is broken.
Checking the result before you keep it
The audio preview plays inside the page, so you can confirm the extraction caught the right track before saving anything. This matters for MKV and MOV files that carry several audio tracks: the tool extracts the default stream, and if a film ships with an English and a French track, the one marked default is the one you get. There is no track selector.
Which bitrate to choose
Bitrate is the only quality dial the tool exposes, and the three values map to clear use cases.
| Bitrate | Roughly per minute | Use it for |
|---|---|---|
| 128 kbps | About 1 MB | Speech, lectures, voice memos, meeting recordings |
| 192 kbps | About 1.4 MB | The default. Music and general listening |
| 320 kbps | About 2.4 MB | Music you will archive or re-encode from later |
Those figures come from the arithmetic of a constant bitrate: 128 kilobits per second is 128 times 60 divided by 8 kilobytes per minute, which is close to 1 MB. Because the encoder writes MP3 at a constant rate, the file size is predictable before you press the button, and the panel confirms it afterwards. For comparison, CD quality PCM at 44.1 kHz in stereo runs at about 10 MB per minute, which is why the same audio as WAV is roughly ten times the size.
What this tool does not do
These are the real edges of a browser-based extraction, and knowing them saves a wasted attempt.
- It cannot cut the audio. There is no start and end trim in this tool. Extract the whole track, then cut it with the audio trimmer, which shows the waveform and exports the section you keep.
- It cannot accept an audio-only file. The drop zone takes video, so an MP3 or WAV cannot be used as input. Use the audio converter for format changes between audio files, including batch conversions.
- There is no volume normalization. A quiet recording stays quiet. The encoder passes the level through untouched.
- No bitrate above 320 kbps and no VBR. MP3 and M4A are written at a constant bitrate from a fixed list of three values.
- Only the default audio track. Files with multiple embedded tracks give you the default one, with no chooser.
- The first run downloads the engine. About 30 MB of FFmpeg code, cached afterwards. It is the cost of encoding locally instead of on someone else's machine.
- Long files are slow. A full length film will take a while, because the work happens on your device rather than on a server with more cores.
When extracting the whole track is the wrong move
A two hour recording that you need ten minutes from is a bad candidate for a straight conversion: you would hold the whole file in memory, convert all of it, and then cut the result. Trimming first is faster at every step. The video trimmer cuts the clip before extraction, and the audio trimmer cuts the extracted file afterwards if you only discovered the useful section late.
The same logic applies to a clip that is mostly silence or a single speaker you need one answer from. Extract, then trim, then keep only what earns its space. It is also worth checking whether the video needs to exist at all afterwards: if the picture was the point, keep the video and use the video compressor instead of throwing the frames away.