Free · 100% offline · by The Treq
Clean up AI music artifacts.
Then master for release.
Shimmer finds the noise AI music tools add, like fizz on cymbals and grain on vocals, removes it, and masters your song for release.
The problem
You generate a track and it sounds great — until you really listen. A thin, fizzy hash rides the cymbals. The vocals wear a glassy sheen. Turn it up on good speakers and it gets worse.
That's a generation artifact — hiss, flicker, and metallic ringing that AI music models leave behind, usually between 5 kHz and 12 kHz. Normal mastering tools treat it as part of your music, so they compress it, brighten it, and make it louder.
What Shimmer does
That order matters. Master first and you just make the fizz louder.
Three released songs, before and after Shimmer. Press play below, then switch between Original, Processed and Removed.
Removed is only what Shimmer took out, turned up so you can hear it.
Any Suno or Udio export. Shimmer listens and finds what the AI added.
Three sliders, already set for your song.
Every stage lights up as your song goes through it.
What's inside
Hit Analyze and Shimmer listens for vocal grain, sibilance, harshness, fixed tones, a dull top end, low-mid mud and low loudness. Quick master then gives you three sliders — Clean-up, Loudness and Tone — already set for your song. Switch to Advanced for every control.
Commercial (−9 LUFS), Balanced (−11) or Streaming (−14), a tone target, and a true-peak limiter. Loudness is one steady gain move — never multiband squash. Then Shimmer checks the file it wrote and tells you if it’s ready to upload.
Bells, shelves, cuts and notches, with presets to start from. Suggested EQ plans a few moves for your song — cuts before boosts — and every move stays editable.
Split any track into vocals, drums, bass, and other, from Fast up to Studio quality. Formant, saturation, doubler, and reverb per stem, with mute/solo/faders. Stems cache per track; your mix saves itself.
Point it at a folder and let it work, file by file. Album mode masters it as one record, so the quiet songs stay quieter than the single.
A plain-language map of every stage your audio passes through. Click a stage to read what it does and jump to its controls. No black box.
The engine
Each fix stays in its own lane. The de-esser works at 4.5–10 kHz, harshness at 2–5 kHz, mud at 200–500 Hz — and only while the problem is there.
A fix runs only when it's on. Every stage has a fixed place in the chain. Turn everything off and the output is your input, bit for bit.
What you preview is what you get. The live preview is the same render as the file you download.
Why it's reliable
Every fix and detector is measured before it ships, and most are classic signal processing. The one trained model — the Shimmer card's de-noise — is small, runs on your CPU, and comes with the app. The same settings give you the same result, every single time.
Your music stays on your computer. Shimmer runs entirely on your machine. No uploads, no accounts, no training on your tracks, no internet needed after setup.
start.bat.If Windows warns about a downloaded file, click More info → Run anyway.
git clone https://github.com/henricksmedia/shimmer.git
cd shimmer
./start.shOr download the latest release, unzip it, and run ./start.sh.
winget install ffmpeg, brew install ffmpeg or apt install ffmpeg. WAV, FLAC and OGG work as they are.Shimmer is free, under the Shimmer License 1.0.