Audio Cutter & Trimmer

Yes. Cuts are made at the AudioBuffer sample level, so the in-point and out-point are precise to within one PCM sample (typically 1/44100 of a second for CD-quality audio). The visual handles in the waveform may snap to the nearest pixel for ease of use, but the underlying sample offsets are exact.

Browsers cannot encode MP3 directly because the format requires licensed codecs. We re-encode via Opus (in WebM) or WAV (uncompressed PCM). Opus at 128 kbps typically matches MP3 at 192 kbps in perceptual quality, and the file is roughly the same size. If you need a true MP3, export as WAV and run it through a desktop encoder like LAME or FFmpeg.

The Web Audio API can decode files up to the browser's memory limit, which on a typical desktop is enough for tracks several hours long. For very large files (over 200 MB), decoding may take a few seconds and the waveform will only render the first portion; trimming still works on the full file but you will need to scrub to find your region.

Not directly in this tool. The cut is a hard splice at the exact sample you choose. If you need a fade, export the trimmed clip and apply a fade-in / fade-out in a desktop audio editor such as Audacity, which is also free.

No. All decoding, trimming, and re-encoding happens locally in your browser using the Web Audio and MediaRecorder APIs. You can verify this by opening your browser DevTools Network tab and watching zero network requests fire during the cut operation.

The waveform uses color to indicate the selected region (between the two handles) versus the unselected regions. Selected samples appear in the accent color and unselected samples in a muted tone so you can clearly see what will be exported.