By ShinobiTools Team · Last updated: July 2026
Hum, hiss, fans, traffic, a laptop droning under a call — CleanGrab pulls the steady background noise out of any voice recording and hands back a clean voice as a 192 kbps MP3 and a lossless WAV. Free, no account, no watermark, no cap.
Clean a recording, free →CleanGrab runs DeepFilterNet3, an open-source speech-enhancement model, on our own hardware. It's trained to tell a human voice apart from the constant noise sitting underneath it, and to strip that noise while leaving the voice natural. The kinds of noise it handles well:
It accepts audio and video files (MP3, WAV, FLAC, M4A, AAC, OGG, Opus, plus MP4, MOV, MKV, WebM and more), up to 60 minutes each, and pulls the voice track straight out of a video if that's what you upload.
Most "my recording sounds bad" problems come down to these two. Hum is the low, steady buzz from mains electricity or a poorly-grounded cable — a single low pitch with overtones stacked above it. Hiss is the higher, airier static from a cheap microphone, a room's own noise floor, or a recording made with the gain turned up too far. Both sit constantly beneath the voice rather than in bursts, which is exactly the pattern DeepFilterNet3 is built to recognise and remove. You don't need to know which one you have — upload the file and the model handles both.
The three strengths are just how hard the model pushes. Start on Balanced; it's right for most files. Use the guide below when it isn't.
Home recordings usually fight a fan or a bit of room hum plus microphone hiss. Balanced clears most of it; reach for Strong only if a heavy fan or AC is clearly audible, since aggressive settings can thin a voice you've worked to record well.
A recorded call picks up more than you'd think: the recording laptop's own fan, traffic through a window, someone typing notes, the hum of an old monitor. Export the recording (video is fine), start on Balanced, and go Strong if the room noise is heavy — but drop back to Gentle if voices start sounding thin, because call audio is often already compressed and doesn't have much headroom to spare.
Live-venue recordings carry room hum, audience shuffling and sometimes wind on an outdoor mic. Balanced is usually the sweet spot here: it preserves the speaker's natural tone, which matters more on a long recording than squeezing out the last bit of ambience.
| Strength | How hard it pushes | Use it when |
|---|---|---|
| Gentle | Up to ~6 dB of reduction | The recording is already fairly clean, or the voice's character matters most |
| Balanced | Up to ~12 dB — the default | Most recordings: removes steady noise without over-processing |
| Strong | No cap on reduction | Heavy background noise; accept a little voice thinning to get rid of it |
Heavy room echo and reverb are a harder, separate problem: CleanGrab reduces light echo, but it will not fully erase a very reverberant room. We'd rather say that plainly than promise something the tool can't do.
CleanGrab is tuned for voice and speech — podcasts, interviews, calls, voice memos, lectures. On music or other non-speech audio the result can be poor, and the tool will tell you when it couldn't detect clear speech instead of handing you a mangled file.
Steady background noise sitting under speech: electrical hum and buzz, hiss and static, fans and air conditioning, traffic and street noise, and keyboard clatter. The voice is kept natural while that layer is pulled out.
Hum is a low, steady electrical buzz at the mains frequency (50 or 60 Hz) and its overtones, usually from a ground loop or a bad cable. Hiss is higher-pitched static from a microphone's own noise floor or a badly-gained recording. CleanGrab removes both.
Yes. Export the recording — video files work too, CleanGrab pulls the audio out — and it removes laptop-fan noise, room hum, traffic and keyboard clatter from the voice track before you share, archive or transcribe it.
Heavy room echo and reverb are a harder, separate problem: CleanGrab reduces light echo, but it will not fully erase a very reverberant room. We'd rather say that plainly than promise something the tool can't do.
Start on Balanced. Go Strong if noise is still audible, or drop to Gentle if the voice starts to sound thin or over-processed. Gentle preserves the voice's character, Strong is the most aggressive.
Free with no account, no watermark and no daily or monthly cap — the ads around the tool cover the hardware. Your upload is deleted the moment it's processed and the cleaned file is wiped within 45 minutes; nothing is stored or used for training. See privacy.
Working with video? CleanGrab cleans the audio inside a video file and hands the video back with the picture untouched — see removing background noise from a video, or the workflow for cleaning up audio for a YouTube video.
Weighing it up against the paid tools? See the honest free Adobe Podcast alternative comparison.
Three situations have their own guide, because each one has causes and fixes this page doesn't cover: background noise in an audio file (what the three strengths change in decibels, and which noise the model simply can't remove), hum and hiss (where they come from in your own setup, and how to kill them at the source in thirty seconds), and a Zoom or Teams recording (why per-speaker tracks beat one mixed file, and what the platform already did to your audio).