100% Free · Version 1.0.0

SoundGeek: clean up a recording, on your own machine

Drop in an interview, a meeting or a voice note and SoundGeek writes a cleaned copy beside it: background noise gone, mains hum gone, levels evened out. It runs entirely on your own computer, so the recording never leaves it.

Windows 10 or 11, 64-bit · self-contained installer · nothing else to install


🆓
100% free
No paid tier, no per-minute limit, no watermark
🔌
Nothing is uploaded
Every recording is read and written on this machine
📦
Self-contained installer
Bundles .NET 8. Nothing to install first
📊
Reports what it changed
In numbers, not in adjectives

Everything a bad recording needs

Three cleanups, from a light touch to a speech model, and you choose which.

🎚️

Levels evened out

Loudness measured to ITU-R BS.1770, the way broadcasters measure it, then brought to a target you pick, with a look-ahead limiter holding the peaks so nothing clips.

🔇

Mains hum removed

SoundGeek measures whether the hum is 50 Hz or 60 Hz rather than assuming, then takes out the fundamental and its harmonics completely.

🌬️

Background noise reduced

Hiss, a fan, traffic, room tone. Spectral subtraction that leaves sustained notes and speech standing rather than flattening everything it finds.

🧠

A speech model for the bad ones

The strongest cleanup separates a voice from the noise around it far better than any filter. Half a megabyte, downloaded when you ask for it, run on this machine.

🪨

Rumble taken out

Desk thumps, wind and traffic under 60 Hz, the noise you feel more than hear, removed before anything else is measured.

📥

Reads nearly anything

WAV, MP3, M4A, FLAC, OGG, Opus, WMA, AAC, and the audio out of MP4, MKV, MOV, AVI and WebM.

📋

A queue, not one file at a time

Add as many recordings as you like and SoundGeek works through them, telling you what it found in each.

🛡️

Your original is never touched

The cleaned copy is written next to the file it came from and named after it. If one is already there, the next is numbered.


See it in action

Every screen, exactly as it looks when you run it.

The Clean screen. Drop recordings in, or choose them.
Three cleanups, from a light touch to the speech model.
A queue of recordings waiting to be cleaned.
What it found and what it changed, in numbers.
What SoundGeek will and will not do, in plain words.
The About box, with the rest of the range.

Who it is for

Anyone with a recording that is worth keeping and hard to listen to.

🎙️

Podcasters and interviewers

A room with a fan in it, a guest on a laptop microphone, a hum you did not hear until you got home. All fixable after the fact.

💼

Anyone recording meetings

A meeting recording that has to be listened to, or fed to a transcriber, is worth two minutes of cleanup first. Nothing about it leaves your machine.

🎓

Students and researchers

Lecture recordings and field interviews, cleaned up without a subscription and without uploading somebody else’s voice to a service.

Requirements

🖥️

Operating system

Windows 10 or later, or Windows 11. 64-bit.

🔒

Rights

Standard user. Administrator is only needed to install it.

📥

Download

About 34 MB. A single self-contained installer.

⚙️

Prerequisites

None. The .NET 8 runtime is bundled. ffmpeg is optional and only needed for formats other than WAV.

🏷️

Version

1.0.0, released 1 September 2026.

📜

Licence

GPL-3.0. Free for everyone, including commercial use.

Frequently Asked Questions

Is it really free? No Pro version, no catch?

Yes. SoundGeek is free to download and use, at home and at work, with no paid tier, no per-minute limit and no watermark on anything it produces.

Does SoundGeek upload my recordings anywhere?

No. Every recording is read, processed and written on your own computer. There is no account and no server. The only thing SoundGeek ever downloads is the speech model, from the Model screen, when you press the button.

Will it overwrite my original recording?

No. The cleaned copy is written next to the file it came from and named after it. The original is only ever read. If a cleaned copy is already there, the next one is numbered rather than replacing it.

Why does the strongest cleanup come back at 16 kHz mono?

Because that is what the speech model works at. It is the right choice for an interview, a meeting or a voice note, and the wrong one for music. SoundGeek says so on the screen next to the option rather than burying it. The other two cleanups give the file back at the sample rate and channel count it went in at.

Why do I need ffmpeg for MP3 and M4A?

SoundGeek does not bundle ffmpeg, because quietly fetching a 90 MB binary you did not ask for is not a reasonable thing for a small app to do. It runs ffmpeg as a separate program if it finds one, and says so plainly if it does not. WAV files work either way.

Can it rescue a really bad recording?

It can make a lot of them usable, and it will tell you in numbers what it changed. Noise reduction cannot put back what was never recorded, so a recording made in a wind tunnel stays a recording made in a wind tunnel.


Version History


Download SoundGeek today

Completely free. Version 1.0.0, no strings attached. Free at home and at work.

Not code-signed, so Windows may warn on first run. Click More info, then Run anyway. SHA256 checksums

Windows may show a “Windows protected your PC” warning the first time you install it. That means SmartScreen has not seen the file before, not that anything is wrong with it. Here is why our builds are unsigned, and how to verify your download against its SHA-256 hash. We will never ask you to switch SmartScreen off.


Support

SoundGeek is free and always will be, at home and at work.

🐛

Found a bug, or want a feature?

Every report gets read, and feature requests genuinely shape what gets built next. The treemap view, the similar-image slider and the scheduled scan mode all exist because somebody asked for them.

Open an issue on GitHub

Buy us a coffee

No tiers, no licence fees, no commercial restriction. If it saved you a subscription, a coffee is welcome. Never expected.

Support us on Ko-fi