Voice Enhancer
Studio-grade clean-up for any voice.
Drop in a podcast, interview, voice-over or video recording and get back clear, even, correctly loud speech. Your audio is never uploaded. Free and unlimited for everyone.
In short: The Voice Enhancer cleans up speech recordings: it removes background noise and echo, evens out the volume and sets the loudness. It runs in your browser, so the recording is not uploaded. It is free and unlimited for everyone, with or without a plan.
How it works
- 1
Drop in your recording
WAV, MP3, M4A, FLAC, OGG, or a video with sound. Each recording can be up to 30 minutes long, on a free account or with a plan.
- 2
Pick a preset
Podcast, interview, noisy place, voice-over or a light touch — or fine-tune every setting yourself.
- 3
Compare and download
Switch between the original and the enhanced version while it plays, see the loudness before and after, then download it as MP3 or WAV.
What you get
Neural noise and echo removal
Powered by DeepFilterNet3, a state-of-the-art open speech-enhancement model, at full 48 kHz quality.
Hum removal
Finds 50 or 60 Hz mains hum and its harmonics automatically and removes it.
Clarity and de-essing
A studio-tone EQ and a de-esser that softens harsh “s” sounds only when they are too strong.
Even volume
Evens out loud and quiet passages without pumping up the background in the pauses.
Broadcast loudness
Sets the loudness to −16 LUFS for podcasts, −14 for YouTube and Spotify, −23 for broadcast, with a true-peak limiter so it never clips.
Before and after
Play the original and the enhanced version side by side, then download 16-bit or 24-bit WAV.
Your audio never leaves your device.
The enhancer runs in your browser. We only count how many times you use it, for your Account page — we never receive the audio, before or after.
Questions
Is it really free?
Yes. The Voice Enhancer is free and unlimited for every signed-in account, with or without a plan. Each recording can be up to 30 minutes long.
Is it the same as Adobe Podcast Enhance Speech?
No. Adobe’s is a proprietary generative model. KUKU Labs uses the open DeepFilterNet3 model plus broadcast-style audio processing. It removes noise and much of the room sound very well, but it won’t rebuild a recording that is badly distorted or very echoey.
Is my audio uploaded?
No. Everything happens in your browser. We only record that a run happened and how long the audio was, so your Account page can show it.
What files can I use?
WAV, MP3, M4A, FLAC and OGG recordings, and videos that contain sound. The result is a mono, 48 kHz WAV.
How long does it take?
It depends on your device. A modern computer processes speech several times faster than real time; a very long file on a phone will be slower, so use a computer for 30-minute files.
Ready when you are.
Text to speech needs no account to try. A free account gives you 5,000 characters a week and unlimited voice enhancing.
Enhance a recording free