Remove Silence from Audio

Silence removal works by measuring the loudness of the audio moment-to-moment and cutting any stretch that stays below a threshold for long enough to count as a pause. This tool uses FFmpeg's silenceremove filter with a threshold around -50 dB, which is quiet enough to treat true dead air and room tone as silence while keeping soft speech and quiet musical passages intact.

Three things get removed: the leading silence before the first sound, the trailing silence after the last sound, and internal gaps longer than a fraction of a second. The audible content itself is never altered — only the quiet regions between and around it are shortened, so the speech or music you keep sounds exactly as it did before.

It is worth being clear about what this tool does and does not do. Silence removal deletes quiet sections of time; it does not clean up hiss, hum, or background noise inside the parts you keep. If a recording has constant background noise, that noise raises the 'silent' level and can even prevent gaps from being detected, because nothing ever drops below the threshold.

For noisy recordings, run a noise-reduction pass first (in a tool like Audacity) and then remove the silence, or record in a quieter space. For clean recordings — a decent microphone in a normal room — automatic silence removal works well out of the box and is the fastest way to tighten the pacing.

Every second of silence you cut is a second less of audio to store, upload, stream, and listen to. For a talky recording — a podcast, a lecture, a long voice memo — pauses and dead air can easily add up to 10-20% of the total runtime. Removing them shrinks the MP3 proportionally and trims the same amount off the listening time.

The editing-time savings are often bigger than the file-size savings. Instead of scrubbing through a recording hunting for gaps to delete by hand, you get a tightened first cut in seconds and only fine-tune from there. For batch workflows — a season of episodes, a folder of interview clips — that adds up fast.

Voice memos, interviews, and meeting recordings are often sensitive, so it matters that they are not uploaded anywhere. This tool runs FFmpeg compiled to WebAssembly entirely inside your browser, on your own device. The audio is read from your disk into your browser's memory, processed there, and offered back as a download — it never touches a server.

Once the page has loaded you can even disconnect from the internet and the tool still works, which is proof there is no upload step. That makes it safe for confidential interviews, private dictation, and client recordings.