Cut the dead air out of podcasts, voice memos, lectures, and language-practice recordings. Detect silent pauses, preview what'll be cut, download the cleaned audio. Runs in your browser — nothing is uploaded.
No file handy?
Drop audio file here
or click to browse — processed in your browser, never uploaded
MP3WAVOGGFLACAACM4A
Your file is never uploaded. All silence detection happens in your browser — nothing is sent to a server. Privacy policy.
—
—
Loading waveform…
KeptWill be cutAdjust the sliders below to see different cuts.
A/B the result before downloading.
-40 dB
0.5 s
0.10 s
0 silent regions detected
New length: — (was —)
—
MP3 320 preserves quality with broad compatibility.
Remove silence from podcasts, voice memos, and lectures — free, private, no sign-up
SnipSound's Silence Remover scans your recording for stretches of quiet and cuts them out. It's built for the post-recording cleanup that every podcaster, lecturer, and language learner ends up doing: shortening a 45-minute raw take into a 32-minute publishable episode without manually finding and deleting every awkward pause in a DAW.
How the detection actually works
The tool walks through your audio in small windows (about 20 milliseconds each) and computes the loudness of each window in decibels. Any window quieter than your threshold is flagged as silent. Adjacent silent windows merge into regions. A region is only removed if it's longer than minimum silence, and the cut leaves padding seconds on each side so the seam doesn't sound spliced. The defaults are tuned for conversational voice; the sliders let you tighten them for a deadpan podcast or loosen them for a lively interview.
What it's great for
🎙 Podcast post-production
Shave 10–25% off a raw episode by removing dead air, ums, and re-take pauses. Pair with the Volume Booster to normalize before publishing.
🎓 Lecture & class recordings
Strip out long whiteboard-writing silences, equipment fiddling, and the long pause between question and answer.
📝 Voice memos & interview transcripts
Tighten an ad-hoc recording before sending to a transcription service, or before sharing with a co-worker.
🗣 ESL & pronunciation practice
Remove the gap between attempts when you're rehearsing the same phrase multiple times. The result lines up nicely with shadowing exercises.
🎧 Audiobook chapter prep
Cut the page-turn pauses and long inter-paragraph breaths from a self-narrated audiobook chapter.
📞 Call recordings & meetings
Compress a 60-min Zoom recording into the meaningful 35 minutes so it's faster to re-listen and share.
Suggested settings
If the defaults aren't matching your material, here are starting points:
-40 / 0.5s / 0.10s
Default (general voice). Works for most podcasts and lectures recorded with a decent microphone in a quiet room.
-35 / 0.4s / 0.08s
Aggressive. Cuts more, including some shorter natural pauses. Use when you specifically want a punchy, fast-paced edit.
-45 / 0.7s / 0.15s
Gentle. Only removes obvious long pauses. Preserves a more natural rhythm. Good for thoughtful interviews and storytelling.
-50 / 1.0s / 0.20s
Quiet rooms. If your recording has very low background noise, you can use a stricter threshold. The longer minimum silence ensures only real pauses get cut.
-30 / 0.5s / 0.10s
Noisy rooms. If there's background hum or fan noise, raise the threshold so quiet speech doesn't get mistaken for silence.
Why use SnipSound's silence remover
100% in your browser. Other tools (Cleanvoice, Descript, Auphonic) upload your file to a server and run cloud DSP. We don't — your audio never leaves your device.
No sign-up, no quota, no watermark. No monthly minutes limit, no "free until you've processed 30 minutes" trick. The whole tool is free, every time.
Adjustable, not magic. Three transparent sliders. You see exactly what'll be cut before you commit. No mysterious AI "smart cut" that you can't tune.
Works on any voice format. MP3, WAV, M4A from your iPhone Voice Memos app, OGG from a browser recorder, FLAC from a portable recorder — all decode and process locally.
How does the silence remover decide what's silent?▼
It scans the audio in short windows (about 20 milliseconds each) and measures the loudness of each window in decibels. Any window quieter than your chosen threshold counts as silent. Threshold is dB relative to digital full scale — a real, clearly-quiet voice recording typically sits between -50 dB and -35 dB during pauses, so the default of -40 dB is a sensible starting point.
What's the difference between threshold, minimum silence, and padding?▼
Threshold is the loudness ceiling for "silent" (lower = stricter, fewer regions detected). Minimum silence is how long a quiet stretch has to be before it counts as a real pause worth removing (e.g. 0.5s skips natural inter-word gaps). Padding is how much of the pause to leave intact at each end so the result doesn't sound chopped — a tiny breath on each side prevents the seam from sounding glued together.
Will this destroy the natural rhythm of speech?▼
Only if you set padding too low and minimum silence too short. The defaults (-40 dB / 0.5s min / 0.1s padding) preserve breaths and beats while cutting the long uhh-pauses that bloat a recording. If the output sounds rushed, raise padding to 0.2s or raise minimum silence to 1.0s.
Does it work on music?▼
It can — but you probably want to use the Audio Trimmer instead. Music typically has dynamics and reverb tails that this tool will flag as silence. It's tuned for voice.
Is my file uploaded to a server?▼
No. All silence detection and audio re-encoding happens in your browser using the Web Audio API. Your audio never leaves your device.
What input formats are supported?▼
Any audio format your browser can decode — MP3, WAV, OGG, FLAC, AAC, M4A, and several others. Output formats include MP3 (320, 256, 192, 128 kbps), WAV, FLAC, OGG, AAC and M4A.
How long can my file be?▼
There's no hard limit — files run only as long as your device's memory allows. A typical laptop handles a 60-minute podcast in a few seconds. A 4-hour lecture may take 30+ seconds to decode and process.
Snip