AntiUpload// browser-resident file tools
ENESFRPTDE
SESSION · —
← Back to home

Descript-style · Two-pass · Audio or video · No watermark

Auto-Cut Silence

Remove dead air from podcasts and videos — silencedetect finds the quiet parts, the rest concats together. Adjustable threshold and padding.

100% freeNo file size limitNo watermarkNo sign-up
  1. 1Pick file
  2. 2Configure
  3. 3Download
Audio or video input. Defaults (-30 dB threshold, 0.5s minimum silence, 0.15s padding) are tuned for typical podcast / voice content. Adjust below if the tool cuts too aggressively or too conservatively.
  • Files never leave your browser — processed entirely on your device
  • No upload, no queue, no waiting for a worker to free up
  • No file-size cap from us — limit is your device's RAM

About Auto-Cut Silence

AntiUpload's Auto-Cut Silence finds the silent stretches in a podcast or video recording and cuts them out, leaving the speech (or whatever non-silent content) intact. It's the Descript / Adobe Podcast Enhance / Riverside silence-removal feature, free, running locally in your browser via FFmpeg's `silencedetect` filter. The two-pass workflow: pass 1 scans the audio for stretches quieter than your threshold (-30 dB default) lasting longer than your minimum (0.5s default), and emits silence_start / silence_end timestamp pairs. Pass 2 inverts those into speech ranges (with a configurable padding so words don't get clipped at cut points), builds a filter_complex graph that trims to each speech range, and concats them back together.

The economics matter: Descript charges $24/month for the editing suite that includes this feature; Adobe Podcast (the closest free competitor) limits to 1 hour/month free tier with intermittent quality issues. Our tool runs locally, has no time cap, and produces predictable output (you control the threshold and padding, not an opaque ML model). The trade-off: we use a simple energy-based silence detector (FFmpeg silencedetect), not the speech-aware detector Descript uses. If you have background music that drops below threshold in places, our tool will cut it; Descript's model knows "there's still music underneath, don't cut." For pure-voice content (podcasts without background music, voicemails, meeting recordings) the simple detector matches the smart detector's behaviour at zero cost.

The threshold (-30 dB default) and minimum-silence-duration (0.5s default) are the two main knobs. Lower threshold (more negative, e.g. -40 dB) cuts only the truly silent parts — safer, conservative. Higher (less negative, e.g. -20 dB) catches quieter ambient noise as "silent" — aggressive cut. The padding (0.15s default) is the buffer of speech kept on each side of every cut so the first and last word of each segment aren't clipped. Works on both audio and video files — for video, the picture stays in sync with the audio cuts because we trim both streams simultaneously and re-encode the result.

How it works

  1. Drop your audio or video fileAccepts every common video container (MP4 / MOV / WebM / MKV / AVI) and every common audio format (MP3 / WAV / M4A / OGG / FLAC / AAC / OPUS). Video stays in sync with audio cuts.
  2. Set silence threshold (dB)-30 dB default works for typical podcast / Zoom voice. -40 dB for very quiet recordings (kid asleep nearby, ambient noise floor needs respecting). -25 dB if your audio is loud and you want aggressive cuts.
  3. Set minimum silence (seconds)0.5s default keeps natural beat-pauses ("uh", thinking time) and cuts only longer dead air. Increase to 1.0s for more conservative cuts. Decrease to 0.3s for aggressive pacing.
  4. Set padding (seconds)0.15s default keeps a small buffer of speech on each side of every cut so words don't clip. Bump to 0.25s if you hear word fragments at cut points. Drop to 0.05s for tighter pacing if your speech is clean.
  5. Click Remove silencesPass 1 scans the audio (~10% of total time). Pass 2 trims + concats the speech segments (~90%). Output preserves the source format for audio inputs; video inputs always output as MP4.

When to use Auto-Cut Silence

Tightening a podcast / interview recording
Most podcasts have ~10-20% dead air (thinking pauses, "um", waiting for the guest to respond). Auto-cut removes the longer pauses (>0.5s) without touching natural rhythm. A 60-minute raw recording typically becomes a 45-50-minute cut.
Cleaning up Zoom / Google Meet recordings
Meeting recordings have 30-50% dead air. Set threshold to -30 dB and minimum to 1.0s to cut the long mute periods between speakers without trimming natural pauses. Saves hours of editing.
Compressing a long lecture for archive
Even paced lecturers have long pauses for writing on a board, answering thinking-out-loud questions, etc. Auto-cut shrinks an hour-long lecture by 15-30% without losing content.
Pre-processing video for auto-subtitles accuracy
Whisper-tiny (our Auto Subtitles tool) hallucinates captions on silent regions ("Thanks for watching", "[music]"). Pre-cutting silences fixes this — Whisper only sees the speech and produces fewer hallucinated lines.
Editing voice memos / dictation into tight clips
iPhone Voice Memos and Android voice recorders capture a lot of "ums" and dead air. Auto-cut tightens the recording to listenable pacing without manually finding and trimming each pause.

Frequently asked questions

How to remove silence from a podcast for free?
Drop your podcast file into our Auto-Cut Silence tool. The default threshold (-30 dB) and minimum-silence (0.5s) are tuned for typical podcast voice. No watermark, no upload, no size cap. Compare to Descript ($24/month for the editing suite that includes this) or Adobe Podcast (limits to 1hr/month free).
What's the difference between this and Descript's silence removal?
Descript uses a speech-aware ML model that understands "this is silent but background music is still playing, don't cut." Our tool uses energy-based detection (FFmpeg silencedetect): anything below the threshold for longer than the minimum is silent, regardless of context. For pure-voice content (podcasts, meetings, voicemails) the difference is negligible. For content with background music that has quiet passages, Descript handles better — but ours is free.
My output sounds choppy — what should I change?
Increase the padding (0.15 → 0.25 or 0.3 seconds). This keeps more speech on each side of every cut, smoothing the edits. Also try increasing the minimum silence duration (0.5 → 1.0 or 1.5 seconds) so only longer dead air gets cut, preserving natural beat-pauses.
Tool says "no silences found" but my file has obvious dead air
Your dead air is louder than the -30 dB threshold — common in noisy environments (kitchen, outdoors, computer-fan-heavy room). Raise the threshold (-30 → -25 dB) so quieter ambient noise counts as "silent." Trial-and-error until the tool finds the pauses you hear.
Can I use this on video to remove dead air from a vlog?
Yes. Drop the video, the tool processes the audio for silence detection and trims both video and audio streams in sync. Output is always MP4 for video inputs. Useful for vlogs, tutorials, screen recordings with pauses.
Will this hurt audio quality?
The output gets re-encoded (video inputs always re-encode to MP4, audio inputs re-encode to their original format). For MP3 / lossy formats this is a generational loss vs source — minimal but not zero. WAV / FLAC inputs round-trip losslessly. If preserving original audio bitrate is critical, work with WAV / FLAC source.
Best free auto-silence-remover online?
AntiUpload Auto-Cut Silence — no watermark, no signup, no time cap, no upload. Compare to Descript ($24/mo), Adobe Podcast (1hr/mo free), Veed ($25/mo to remove watermark). All of those upload your file; ours runs in your browser.

Related tools

Auto SubtitlesGenerate .srt or .vtt captions from speech using Whisper-tiny — runs entirely in your browser, downloads a 75 MB model on first useVideo TrimmerCut a clip out of a video — fast stream-copy by default, with optional frame-accurate re-encode and fade-in / fade-outAudio TrimmerCut a clip out of an MP3, WAV, M4A, OGG, FLAC, AAC, or OPUS — fast and lossless, no re-encodeChange Audio VolumeMake an audio file louder or quieter — multipliers from 0.25× to 3×, output to MP3 / WAV / M4A / OGG / FLAC