Add audio to a video: music, a voice-over, or a new sound.
Drop a video and a song or recording. Replace the sound, mix the two, or let the music duck under the speech by itself. The picture is copied as it is, so nothing is lost and a long clip takes seconds. Nothing is uploaded.
drop the video and the audio
MP4, MOV, MKV or WebM, plus an MP3, WAV, M4A or any other audio. Both at once, or one and then the other.
At a glance
| Takes | A video (MP4, MOV, MKV, WebM, with or without sound) and any audio the site reads (MP3, WAV, M4A, FLAC, OGG, AIFF, WMA, voice notes) |
|---|---|
| Modes | Duck the music under speech, mix the two, or replace the sound |
| Ducking | 6, 12, 18 or 24 dB under speech, back to full level in the pauses; measured to within 1 dB of the setting |
| Fitting | Start offset, a start point inside the track, looping with crossfaded seams, a fade-out of up to 5 s |
| Picture | Copied frame for frame, never re-encoded; MOV and MKV keep their container, the rest become MP4 |
| Sound out | AAC 192 kbps at the video's own sample rate, limited to -1 dBTP so it cannot clip |
| Uploads your files | No. No watermark, no account |
How to add music to a video
Drop the video and the song on the box above, together or one at a time. The page keeps the video's own sound and turns the music down whenever someone talks, which is what most clips want: a vlog, a product demo, a talking-head reel, a screen recording with narration. Set how loud the music is with "New track level", press Make the video, and the file downloads. The first run fetches the video codec once (about 30 MB, kept by the browser), so later runs start straight away.
For a clip with no speech at all, such as a drone shot, a timelapse or a game capture, choose "Replace it with the new track" and the music becomes the whole soundtrack. "Mixed with the new track" keeps both at the levels you set and does nothing clever, which suits a live recording where the room sound matters.
Ducking: music that gets out of the way
Editors call it ducking or sidechain compression: the music dips while a voice is on and comes back in the gaps. Here it is worked out from the video's own sound. The page listens to the speech band (roughly 250 Hz to 3.5 kHz), learns that recording's background level, and treats anything 10 dB above it as talk, so it works the same on a phone clip in a kitchen and a lavalier in a quiet room. The music falls in about 60 ms when a sentence starts, holds through the short gaps between words, and comes back over half a second, slowly enough that it doesn't pump.
12 dB is the default because it keeps a voice clear over most songs without the music vanishing. Pick 6 for a subtle bed under a voice that is already loud, or 18 to 24 when the track is busy or has its own vocals competing with the speaker. The check that ships with this page measures the music at the chosen depth to within 1 dB under speech, and back at full level in the pauses.
Fitting a song to the clip
The finished file is always exactly as long as the video. A song that runs longer is cut at the end, and the 3-second fade-out stops it ending mid-note. A track that is shorter loops: each repeat is crossfaded into the next over a second and a half, so there is no click or gap at the join, though a song with a clear ending will still be heard starting again. "Starts at" delays the music, for a title card or an intro line before it comes in. "From" skips into the track, to start a song at its chorus rather than a long intro.
Why the picture is not touched
Most online editors decode and re-encode the whole video to change its sound, which costs quality and time and is where watermarks get added. Here only the sound is rebuilt. The video stream is copied into the new file packet for packet, so a 4K clip comes out exactly as sharp as it went in, every frame is kept, and the work takes about as long as the audio does, seconds rather than minutes. The original sound is not moved against the picture: the check measures 0 ms of drift.
The new sound is AAC at 192 kbps at the video's sample rate, the format every phone, browser, YouTube, Instagram and TikTok expects. MOV from an iPhone stays a MOV and MKV stays MKV; WebM becomes MP4, because WebM needs Opus audio and this page cannot write Opus.
What it can't do
There is one new track per run, so to add both a voice-over and music, put the voice-over on first and then run the result through again with the music, ducking set. It does not remove the music already in a video while keeping the voices; the vocal remover separates a voice from its backing first. And adding a song does not license it: YouTube's Content ID and Instagram match the sound, and a commercial track will be claimed or muted whatever the picture is. Use music you own or royalty-free tracks.
FAQ
How do I add background music to a video for free?
Drop the video and the song above, keep "duck the music under speech" so voices stay clear, set the music level and press Make the video. There is no watermark, no account and nothing uploaded.
How do I replace the audio in a video?
Choose "Replace it with the new track". The video's own sound is dropped and the new file becomes the soundtrack, looped or cut to the video's length. The picture is copied untouched.
Can I add a voice-over to a video?
Yes. Record it with the voice recorder or make it with text to speech, then drop it here with the video. Choose "Replace" for narration alone, or "mixed with the new track" to keep the original sound underneath.
Does it lower the video quality?
No. Only the sound is rebuilt; the video stream is copied into the new file without being decoded or encoded, so every frame is exactly as it was.
What if the song is shorter than the video?
With "Loop" ticked it repeats, crossfaded at each join so there is no click or gap. Untick it and the video goes quiet when the song ends.
Does it work on an iPhone video?
Yes. An iPhone MOV stays a MOV with the new sound, and it plays in Photos, on Instagram and on TikTok. On the phone itself, very long 4K clips can run out of memory; a laptop handles them easily.
Will YouTube or Instagram accept a song I add?
The file will upload, but copyright is matched on the sound: a commercial song will usually be claimed or muted. Use music you have rights to, or royalty-free tracks.