Voice Activity Detector
AIRuns On Your DeviceFind exactly where speech starts and stops in any audio file — Silero VAD runs in your browser, works for every language, no upload.
This tool runs Silero VAD (~2 MB, downloaded once and cached, works for any language including Mongolian) over your audio file and reports exactly where speech starts and stops. The file never leaves your device.
Typical uses: finding talk time in a podcast or meeting recording, locating segments to cut, or exporting a silence-stripped version.
Select an audio file
Supports MP3, WAV, M4A, FLAC, OGG (Max 200MB)
Voice Activity Detector — Where Does the Talking Start?
Upload an audio file and Silero VAD — a genuine neural voice-activity model trained on 6000+ languages — maps every speech segment with timestamps, entirely inside your browser. The file never leaves your device, which matters for meeting recordings, interviews, and voice notes you cannot upload anywhere.
Get the total talk-time ratio, a visual speech timeline, copyable cut points for editing, and a one-click silence-stripped WAV export. The ~2 MB model downloads once and is cached; after that it works offline.
Frequently Asked Questions
Everything you need to know about using Voice Activity Detector safely and efficiently.
Related Online Utilities
Speech to Text (In-Browser AI)
Transcribe microphone audio or audio files to text in English, Spanish, French, Portuguese, or Japanese — AI that runs in your browser, no uploads.
Audio Trimmer & Cutter
Cut MP3, WAV, M4A, FLAC to an exact segment — trim, fade, and export MP3 or WAV in your browser.
BPM & Tempo Detector
Find the tempo of any song or loop — onset-energy autocorrelation with confidence score and beat grid, fully in-browser.