Background Audio Remover
Remove background noise and music from video and audio, keep the voice
About Background Audio Remover
Background Audio Remover is a web tool that takes unwanted sound out of a recording and leaves the voice. It handles two jobs. Remove noise, the default, deals with wind, electrical hum, fans and air conditioning, traffic and room hiss. Remove music takes out background music under speech, and a Keep other sounds slider decides how much of the remaining sound, such as footsteps, traffic or applause, stays in. Both are done with speech models that have learned what voices sound like, keep them and turn everything else down.
Its main difference from most online cleaners is where the work happens. By default the file is processed in your own browser and is not uploaded, with no account, no watermark and no daily limit. On a computer it accepts files up to 3 hours and 4 GB (60 minutes for Remove music). It also works on the video itself: you drop an MP4, MOV, WebM or MKV and get the same video back with clean sound, with the picture copied without re-encoding. The about page lists the open-source parts it is built on, including the DeepFilterNet3 noise model compiled to WebAssembly, the Bandit v2 speech and music separation model run with ONNX Runtime Web, and Mediabunny for reading and writing media files. The site is run by Deerlight Labs, and paid cloud processing is optional for phones or for faster music removal.
Open the homepage and choose what should go: Remove noise or Remove music. Then drop a video or audio file, or try the built-in 26-second sample. Processing starts on its own. Your browser downloads the noise model once, about 12 MB, and caches it; the larger music model, about 54 MB, is only fetched when you run Remove music on your device.
While the file is cleaned you see a progress bar and the time left. Remove noise usually runs faster than the file plays, while Remove music takes about 2 minutes or more per minute of sound. Long files are split into pieces and processed on several CPU cores, then blended back together.
When it finishes, switch between Original and Clean while the file plays, adjust the Keep other sounds slider if you used Remove music, and download the same file with clean sound. The cleaned audio is encoded at the same bitrate as the original and put back into the same container.
If you prefer, choose Cloud instead. For Remove music, the same model, settings and blending code run on the operator's own processing server, every file gets a free 30-second preview, and the original is deleted right after processing.
- •Free local noise removal - Wind, hum, fans, traffic and room noise are removed in the browser with no upload, sign-up or watermark, on files up to 3 hours and 4 GB on a computer.
- •Music removal with a mix slider - Remove music separates speech, music and other sounds. Set Keep other sounds to 0 for the voice alone or up to 100 to keep ambience. Singing in a background song is treated as music and removed too.
- •Video in, video out - MP4, MOV, WebM and MKV come back as the same video with clean sound, and the picture is copied rather than re-encoded. MP3, WAV, M4A, AAC, FLAC, OGG and OPUS audio also work.
- •Original and Clean comparison - Switch between the two versions during playback before you download, and hear demos on the homepage built from public-domain speech with generated noise.
- •Optional cloud processing - Cloud runs the same pipeline on a server for phones or slower machines, with free 30-second previews and 3 free cloud minutes for new accounts. Failed cloud jobs return their minutes automatically.
- •Open about its parts and limits - The about page publishes the models and libraries used with their licenses, and the site states what it does not do: it does not remove echo or reverb, and other people talking are kept in every mode.
Background Audio Remover is aimed at anyone with a spoken recording spoiled by its surroundings. Vloggers and people filming outdoors can take wind and traffic out of interviews, drone and action-camera clips. Podcasters and lecturers can clean hum and hiss from voice recordings, and people on video calls can tidy a recording before sharing it.
It is a good fit for creators who would rather not upload private footage to a server, since the default mode never leaves the device, and for anyone who needs the video back rather than a separate audio file. The Remove music mode helps when someone is talking over background music and only the speech, or the speech plus ambience, is wanted.
Occasional users can stay on the free local mode or buy a one-time pack of cloud minutes; people who process files every week can take the Pro plan for unlimited cloud processing. It is meant for speech, not for splitting a song into vocals and instrumental.
Pricing
$0 - $14.99/mo
- Free$0/mo
- Starter pack$4.99 once
- Big pack$19.99 once
- Pro$14.99/mo
From the vendor pricing page, 2026-11-05
