Audio Trimmer

'Trim' and 'cut' describe the same operation — picking a start time and an end time, and keeping just the audio between them. The difference is intent. 'Trimming' usually means cleaning up the edges of a recording: removing the silence at the start, the dead air at the end, the mistakes around the real content. 'Cutting' usually means extracting a piece from somewhere in the middle: the chorus of a song for a ringtone, the quotable line from a podcast for social. The tool handles both — pick the framing that matches what you are doing. The button labels and copy on this page lean trim because that is what the search intent is for 'audio trimmer'.

Frame-accurate to the audio sample. The trim points can land at any timestamp you specify, with hundredth-of-a-second precision in the time inputs. For lossy sources (MP3, AAC, OGG, M4A), the output is re-encoded at the original bitrate so the cut can be sample-accurate rather than rounded to a codec frame boundary. For lossless sources (WAV, FLAC, AIFF), the re-encode is bit-identical to the original — no quality change at all. In practice, sub-second precision is more than enough for trimming silence. If you find yourself needing millisecond precision regularly, you are probably doing music editing and a DAW like Reaper or Audacity is the right tool.

The trimmer works on every common audio format: MP3, WAV, M4A (AAC and ALAC), FLAC, OGG (Vorbis and Opus), AAC, Opus, AIFF, and WMA. Plus video files (MP4, MOV) where we extract the audio track and trim that. The output format matches the input — drop in an MP3, get an MP3 out; drop in a WAV, get a WAV out. If you want to change format and trim at the same time, do the trim first and run the result through one of the format conversion tools (/wav-to-mp3, /m4a-to-mp3, etc). Doing it in two passes keeps the operations separate and easier to debug if something looks off in the result.

Other tools try to detect silence and trim it automatically. We do not, on purpose. Silence detection is a guessing game with no universal right answer — what counts as silence depends on the noise floor of the recording, the gain staging, and the user's intent. A speaker who pauses for emphasis sounds like silence to a detector but is not silence to the listener. The breath you take before a punchline reads as silence to the algorithm but is the punchline's setup. Auto-trim guesses wrong often enough that the time saved is offset by the time spent fixing or redoing trims. Manual trimming with a waveform takes about 10 seconds and lands the trim where you actually want it, every time.

The trimmer runs entirely in your browser using FFmpeg compiled to WebAssembly. Your audio is never uploaded. There is no server-side processing, no temporary file storage, no logs of what you trimmed. This matters if you are trimming voice memos with personal details, client recordings under NDA, leaked or pre-release material, sensitive interviews, or anything else you would not hand to a third-party server. After the first page load, the tool works offline — disconnect from Wi-Fi and trim the file with no network access at all.