Search UGESI

↑ ↓ to move, Enter to open, Esc to close Browse all tools

Remove background noise from audio

Upload a recording and get the voice on its own, without the street, the hum or the room.

  • Separates the voice from street, hum and room
  • MP3, WAV or M4A, up to 30 MB
  • Recordings up to 10 minutes
  • 7 credits per track, no subscription
Made with AI on Ugesi: a microphone by a rainy street window with a noisy sound wave turning into a clean one.
Example result
Audio file
Click or drop a file MP3, WAV or M4A, up to 30 MB.
Cost of this generation
7 credits / track
7 credits

New accounts get 100 free credits for fast image models. Create an account

  • No subscription needed
  • Rate shown before you generate
  • Your inputs are never used for training

Create a free account to generate. New accounts get 100 credits for fast image models once the email is confirmed, valid 30 days. Create account →

How to use Audio Noise Remover

  1. Upload the original recording

    Under "Audio file", add the least processed version you have. The original WAV is better than a re-exported MP3. MP3, WAV or M4A up to 30 MB. Recordings longer than 10 minutes are rejected, so trim long files to the part you need first.

  2. Clean the audio

    There is nothing to set: the tool works from the file alone. Press "Clean the audio" and the voice is separated from music, traffic and room noise. It comes back as one audio file.

  3. Listen on headphones and a phone

    Play the result on headphones and on a phone speaker. Where noise sat right on top of the voice, expect faint artifacts or a thinner tone. If it sounds watery, start again from a cleaner, less compressed original. The source file is the only thing you can change.

What a real job costs here

Runs on this page cost 7 credits. At the best pack rate that is about $0.0745.

PackCredits Runs with Stem Splitter
$9750 107
$191,650 235
$393,500 500
$797,300 1,042
$14914,000 2,000

What this model is good at and where it stops

Stem Splitter

Separation beats filtering

Classic noise reduction carves frequencies out and takes the voice with them. Stem separation pulls the voice out as its own track instead, which is why it sounds natural.

What comes back

One audio file: the voice separated from the background. If it sounds too dry, mix a little room tone back in your editor.

Where it helps

Interviews recorded in cafes, phone calls, podcasts with a fan running. Any recording where the moment was right and the room was wrong.

What noise removal takes with it

Every removal is a trade. Separating out room hiss also thins the top of the voice. That is why over-processed audio sounds like it is underwater. There is no strength setting here, so the cleaner and less compressed the original, the less the voice loses. Listen on headphones and on a phone speaker before you use it.

Frequently asked questions

Will it remove music behind speech?

Yes, that is exactly what separation does. Heavy overlap leaves faint artifacts.

What files work?

MP3, WAV and M4A up to 30 MB.

Is it the same as the stem splitter?

Same engine, different intent. The splitter is for music, this page is for cleaning a voice recording.

What does cleaning one recording cost?

7 credits per track. The exact total appears above the button before you generate.

The best AI models and tools, without another subscription

Pay only for what you generate. New accounts start with free credits, no card needed.

Start for free