Remove Silence
Trim leading and trailing silence from recordings automatically with adjustable threshold.
100% in your browser — files never leave your device
Automatic silence trimming for cleaner recordings
Recordings often start and end with silence: time before you begin speaking, or after you stop. This is common in field recordings, voice memos, and podcast raw takes. Trimming that silence manually requires locating the exact sample where audio begins and ends. This tool automates that step.
The tool uses FFmpeg's silenceremove filter to detect where audio crosses the silence threshold, then trims the file to the active audio region. The threshold (in dB) controls what counts as silence, and the minimum duration controls the shortest silence worth removing. For most clean voice recordings, the defaults (-45 dB, 0.5 seconds) work without adjustment.
Processing runs entirely in your browser with FFmpeg.wasm. No upload, no account, no size limits beyond your device's available memory.
Frequently asked questions
What does the silence threshold control?
The threshold sets the dB level below which audio is considered silence. A value of -45 dB means any audio quieter than -45 dBFS is treated as silence. Lower (more negative) values are more conservative — only removing very quiet sections. Higher values (e.g., -30 dB) are more aggressive and may remove low-level background noise along with actual silence.
What does the minimum silence duration control?
This is the shortest silence that will be removed. At 0.5 seconds, brief pauses shorter than half a second are kept intact — only longer gaps are trimmed. Increasing this value means only longer silences get removed, which is useful for preserving natural speech rhythm.
Does this remove silence from the middle of the recording as well?
This tool removes leading and trailing silence. It detects the first moment audio exceeds the threshold (trimming the silent start) and the last moment audio exceeds the threshold (trimming the silent end). Internal pauses within the audio are not removed, which preserves natural speech pacing.
Is my audio uploaded anywhere?
No. All processing runs in your browser using FFmpeg.wasm — a WebAssembly build of FFmpeg. The WASM binary (about 25 MB) downloads from a CDN on first use and is then cached. Your audio never leaves your device.
Why does the first run take a moment to start?
FFmpeg.wasm (about 25 MB) downloads on first use and is cached by your browser. After that, subsequent operations start immediately without any additional download.