Add Audio to Video

Add or replace a video's soundtrack with an audio file, right in your browser. Free and private, with no upload, no watermark and no server.

🌐 Español

Drop your files here (.mp4, .mov, .mkv, .webm, .avi, .mp3, .wav, .m4a, .ogg, .flac)

🔒 Private by design: your files are processed locally in your browser and never uploaded to any server.

Written and maintained by Last updated

One dropzone, two roles, sorted by extension

Most tools that combine two files give you two separate upload boxes. This one has a single box that accepts both lists at once: the video containers .mp4, .mov, .mkv, .webm and .avi, and the audio formats .mp3, .wav, .m4a, .ogg and .flac. Which file is the picture and which is the soundtrack is decided by extension, not by the order you added them, so dropping both together works exactly as well as adding them one at a time.

The rule is strict in one direction only: exactly one of each. If the sort finds two videos, or no audio file, the run stops before FFmpeg is invoked. This is worth knowing because the failure appears as a generic message, and the actual diagnosis is usually visible at a glance in the file list, which stays on screen after a failed run.

Replace and Mix build two different command lines

The replace mode is the straightforward case: the source’s own audio, if it had any, is not mapped into the output at all, and your uploaded track is used as the entire soundtrack. That is what you want for a silent time-lapse, a screen capture you are narrating, or a clip whose original sound you have decided against.

The mix mode keeps both. The two audio streams are fed through a blend filter that turns each of them down by the same factor as it sums them, which is why the result does not distort the way a naive addition of two full-level tracks would. Use it when the on-camera sound matters, dialogue, room tone, game audio, and the new file is a bed underneath it rather than a replacement.

There is a third case the tool handles for you. If the source video carries no audio stream at all, which is normal for exports from screen recorders and time-lapse apps, Mix has nothing to blend with. Rather than fail or produce silence, the run detects the missing stream during a quick probe of the video and quietly builds the Replace command instead. Choosing the wrong mode for a silent clip therefore costs you nothing.

Putting a track over a clip

  1. Drop the video and the audio file onto the box above together, or use Choose files and pick them in either order. They appear as two entries in the file list.
  2. Set Audio mode to Replace or Mix.
  3. Press Add Audio to Video and leave the tab open while the progress bar runs.
  4. Download the result. It takes the video’s own name with -with-audio.mp4 on the end, so it will not overwrite your source.

The copy attempt comes first, and the fallback is silent

Video is expensive to re-encode and lossy every time you do it, so this page tries not to. The first attempt copies the video stream into the new MP4 untouched, changing only what audio sits alongside it. When that works, the picture in the output is identical to the picture you started with and the run is quick, because only the audio needed encoding.

Some sources cannot be copied into an MP4 at all: VP8 or VP9 inside a WEBM, and various older codecs inside an AVI. In that case the first attempt fails, is caught, and a second run encodes the picture with libx264 at a quality-focused setting. You are not told which path ran. The two tells are the time it took and the file size, so if you recorded something with Screen Recorder, which normally writes WEBM, expect the longer path here.

Whichever path runs, the audio side is always re-encoded to AAC at 192 kbps, regardless of whether you brought an MP3, a WAV or a FLAC, and the container is always MP4. That is a deliberate simplification: MP4 with AAC is the combination that plays everywhere without you having to think about which container can legally hold which codec.

Nothing here stretches, loops or lines up your audio

The output ends when the first of the two streams runs out. If your track is shorter than the video, the tail of the video is cut off; if it is longer, the extra audio is discarded when the picture ends. Neither file is looped to fill the gap and neither is time-stretched to match. This surprises people who expect a music track to repeat under a long clip, and it is the single most common reason a result is not the length someone expected.

The fix is to make the two lengths agree before you combine them. For an MP3 soundtrack, Cut MP3 trims it to the length you need without re-encoding. For the picture side, Trim Video handles MP4, MOV, M4V and WEBM sources. And if several clips are going to become one piece, join them with Merge Videos before you score the result rather than after: that page only keeps audio when every single clip it is given has an audio stream, so merging a scored clip with a silent one leaves you with no sound at all.

One more sequencing note: since Replace already drops the original audio, running a clip through Remove Audio from Video beforehand is redundant. That page is for when a silent clip is the actual end product.

Frequently asked questions

I dropped two videos by mistake and the run failed. What happened?

Your files are sorted by extension into exactly one video role and exactly one audio role before FFmpeg is started. Two videos, two audio files, or a missing one of either stops the run at that point. What the page shows is its single generic failure message rather than a note about which role was ambiguous, so check the file list sitting above the button. It should hold exactly two entries, one video container and one audio format.

Can I make the music quieter than the dialogue in Mix mode?

There is no volume control on this page. The blend runs with its default normalisation, which scales both inputs down by the same factor so their sum cannot clip, and the balance you end up with is whatever the two files already had relative to each other. If the music needs to sit under speech, lower its level in an audio editor before bringing it here.

Does the new audio always start at the very beginning of the video?

Yes. Both inputs are mapped from their own first sample with no offset control exposed, so the soundtrack begins at frame one every time. Delaying it means adding the silence to the head of the audio file itself somewhere else first, then combining the padded version here.

Why did my WEBM screen recording come back re-encoded when my MP4 did not?

The video is copied across untouched when the container will take it, and the copy attempt is what decides. VP8 and VP9, the codecs a browser screen recording normally produces, cannot be copied into an MP4 container, so that attempt fails and the fallback re-encode runs instead. An H.264 MP4 usually copies straight through, which is why it finishes far faster and the picture comes back bit for bit identical.

What happens to extra audio tracks or subtitles the source video already had?

They are not carried over. Exactly one video stream and one audio stream are mapped into the output, so a second language track, a commentary track, a subtitle track or chapter data present in the source is absent from the result. The output is always a single-video, single-audio MP4, which is the shape almost every platform and player expects.

Is the finished file ready to upload straight to a platform?

In most cases yes. It is an MP4 carrying AAC audio at 192 kbps with the index moved to the front of the file, which lets a player start before the whole download has arrived. The picture is either your original stream untouched or an H.264 re-encode at a quality-focused setting, depending on which path the run took. Nothing is watermarked.

Related tools