Put a soundtrack under a clip. Load the video, add an audio file, choose whether it replaces the original sound or sits on top of it, then set where it starts and how loud it is. The export copies the picture as it is and encodes only the new soundtrack, so a long video takes about as long as reading the file.
Replace or mix
Replace drops the original soundtrack. A screen recording with keyboard noise gets a music bed; a clip shot on a phone gets a narration recorded separately.
Mix sums the two. Both levels are adjustable, so a voice-over can sit over the original at a lower level, or a music bed can run quietly under an interview. The original level slider only appears in mix mode, and only when the video has a soundtrack to keep.
A video with no audio track behaves the same in both modes: the added audio is the only sound.
Placing the audio
Starts at is the second of the video where the added audio begins. Play the preview to the moment, then press Start at playhead. Before that moment the added track is silent, so a title card can run without music.
A negative start skips into the audio instead. At -30 the export begins thirty seconds into the track, which is how a song starts on its chorus rather than its intro.
Repeat to the end of the video loops a short track until the picture ends. Without it the added audio stops at its own end, and what follows is silence in replace mode or the original alone in mix mode.
Fade out lowers the added audio to silence over the last seconds of the video, whether the track was looped or not.
Levels
Each level runs from 0% to 200%. The sum of the two tracks is clipped at full scale, so two loud sources at 100% each will distort where they coincide. Under speech, an original at 30% to 50% leaves the voice clear. A level above 100% raises a quiet recording, at the cost of clipping its own peaks.
The preview applies levels up to 100%; anything above is heard only in the export.
Output
The output keeps the source’s container, and the video stream is copied without re-encoding whenever that container can hold its codec, which for MP4, MOV, WebM, and MKV sources is nearly always. The new soundtrack is encoded as AAC in MP4 and MOV and as Opus in WebM and MKV, at the higher of the two sources’ sample rates and at most stereo. The result under the controls reports which.
The export is a single pass over the picture and both audio sources, and the output streams to browser storage, so a file larger than memory still works.