Audio Splitter

The honest version: this is the same single-segment widget as the cutter, framed for split workflows. To split one file into N pieces, you run the tool N times — once per piece — with different start/end ranges each pass. We do not auto-split at silence, we do not output multiple files in one click, we do not have a multi-marker timeline. That sounds like a limitation but in practice it is a feature: every split is precise to the sample, every output file is named with the range you chose, and there is no risk of the auto-splitter putting a break in the wrong place. If you split a 5-chapter audiobook, you make 5 passes and you know exactly where each chapter starts and ends. The trade-off is time — for a 50-segment file you would want a desktop tool instead.

The most common feature request for an audio splitter is 'auto-split when you detect silence'. We deliberately do not do this. Silence detection is a guessing game with no universal right answer. The threshold that splits a podcast cleanly between guests will mis-split a music track between verses. The threshold that catches the gap between songs in a DJ mix will incorrectly split during a song's quiet bridge. Auto-split is right about 80% of the time, which means 20% of the splits land in the wrong place and need to be redone manually. Manual splitting with a waveform takes about 15 seconds per cut and lands the split exactly where you want it, every time. For most users, the time saved on the wrong-cases offsets the time saved on the right-cases.

Running the tool once per segment is slower than 'click split and walk away', but it gives you precise control. For each segment you preview the start and end before downloading, so the segment that ends up on disk is exactly the segment you wanted. With auto-splitters, you usually end up listening to each output file afterward to check if the splits landed correctly anyway — and then re-doing the ones that did not. The manual workflow front-loads that work into the cutting pass, where it is faster to fix.

Three cases where a desktop tool wins. First, batch splitting at silence detection across many files: Audacity's 'Sound Finder' or 'Truncate Silence' effects, or a CLI tool like mp3splt with its silence detection mode, beat manual splitting at scale. Second, splitting MP3 files at frame boundaries with zero re-encode: mp3DirectCut on Windows is the tool of choice — it edits at the MP3 frame level so the output is bit-identical to the source within the segment, but it limits cuts to roughly 26ms boundaries. Third, splitting based on cue sheet timestamps (CD rips, DJ mix releases): tools like CUE Splitter or Foobar2000 read the cue file and split accordingly. For a one-off split of one file into a handful of segments, this browser tool is faster than spinning up any of the above. For ongoing batch work, install the right desktop tool.

All splitting runs in your browser using FFmpeg WebAssembly. The audio file is never uploaded. There is no server-side processing and no temporary storage. We have no logs of what you split. This matters because the use cases for splitting (audiobooks you bought, lecture recordings, podcast episodes you produce, DJ mixes you mixed, voice memos with personal details) often involve audio you do not want sitting on a stranger's server even briefly. Disconnect from the internet after the page loads and the splitter still works — proof that there is no server in the loop.