Add audio to video online

A raw video clip from a phone, screen recorder, or camera often has the wrong audio: wind noise from outdoor recording, the fan hum of a laptop, keyboard clicks in a screen recording, or simply no audio at all. Replacing or layering in a clean track is one of the most common finishing steps before sharing any video content.

The use cases span from professional (adding a licensed music track to a product demo) to practical (adding a voiceover to a silent screen recording, mixing background music at a lower volume under a narration). Both require combining two files without re-encoding the video unnecessarily.

The tool has two modes that produce different FFmpeg arguments. In replace mode, the video stream is copied directly with -c:v copy, meaning the video is not re-encoded at all: no quality loss, no processing time beyond the audio. The new audio is mapped with -map 0:v:0 -map 1:a:0 and -shortest trims the output to whichever stream ends first. If the audio file is shorter than the video and loop is enabled, -stream_loop -1 is added before the audio input to loop it continuously.

In mix mode, the tool uses a filter_complex with the amix filter to blend both audio tracks at independent volumes. The volume of each stream is controlled by volume= filters applied before mixing: [0:a]volume=X[a1];[1:a]volume=Y[a2];[a1][a2]amix=inputs=2:duration=first[aout]. In mix mode, the video is also stream-copied (-c:v copy) to avoid re-encoding.

Use replace mode when you want the new audio to be the only audio in the output. Use mix mode when you want both the original audio and new audio to coexist, for example when adding background music under a narration that was recorded with the video.

In mix mode, set the original audio volume to 100 and the new audio to 30 to 40 for a standard background music balance. If you are mixing a narration over music, set the music to 20 to 30 and keep narration at 100.

If the audio file is longer than the video in replace mode, -shortest clips it at the video end. If the audio is shorter and you do not enable loop, the output will have silence after the audio ends.

Video editors handle audio replacement and mixing naturally in a timeline view, but require a project setup and a render pass. Command-line FFmpeg produces the exact same output with a single command, but the amix filter syntax and the stream mapping flags are easy to get wrong. This tool picks the correct argument set for each mode automatically and handles the loop edge case without any flag lookup.