Audio Enhancer
Clean up a voice or music recording — normalize loudness to a consistent level and reduce steady background noise. Runs entirely in your browser; your file never leaves your device.
Everything runs locally in your browser. Your audio is never uploaded. The first run downloads the engine (~25 MB); after that it is cached.
About this tool
The audio enhancer takes a recording that sounds uneven or hissy and makes it more consistent to listen to. It runs two main processes: loudness normalization, which pulls quiet and loud passages toward a single reference level, and steady-noise reduction, which lowers a constant background floor such as tape hiss, mains hum, or a fan running in the room. An optional high-pass filter removes low-frequency rumble from handling and air handling systems.
To be clear about what this is: it is filter-based cleanup built on ffmpeg audio filters, not AI restoration. It improves consistency and reduces steady background noise, but it will not remove loud one-off sounds like a cough or a door slam, and it cannot rebuild audio that is already badly clipped or distorted. It is well suited to voice recordings, meetings, and podcasts, where a stable level and a quieter noise floor make the difference. Nothing is uploaded — a WebAssembly build of ffmpeg does the work on your own machine, which matters for private calls and interviews.
How to use it
Pick a file, choose the cleanup steps, and export:
- Choose an audio or video file, or drag it onto the drop area.
- Leave loudness normalization and noise reduction on for most spoken-word recordings; enable rumble removal if you hear low-frequency thumps.
- Select an output format — MP3 for a small shareable file, WAV when you want a lossless copy for further editing.
- Select Enhance, wait for the progress bar, then download the result.
Common questions
Does my audio get uploaded to a server?
No. The cleanup runs entirely in your browser using a WebAssembly build of ffmpeg. Your file is read locally and never leaves your device, which is why the first run downloads the engine but later runs are quick.
Will this remove every kind of noise?
No. This is filter-based cleanup, not AI restoration. It normalizes loudness and reduces steady background noise such as hiss, hum, and fan or air-conditioner sound. It will not remove loud one-off sounds like a door slam or a cough, and it cannot fix audio that is already badly clipped or distorted.
What is it best suited for?
Spoken-word recordings: voice memos, meetings, interviews, and podcasts. Loudness normalization gives a consistent level across a recording, and noise reduction cleans up a constant room-noise floor. Music can benefit from loudness normalization, but aggressive noise reduction may soften detail.
Need to change formats first, or pull audio out of a video? Try the audio format converter. For heavier media and automation work, see our development and automation service.