Audio Cutter

Trim and cut audio by setting precise start and end points, with the waveform controls to place them accurately and private on-device processing throughout.

• A cut that does not re-encode — keep the source format and change nothing else, and the original audio is copied straight through
• Keep or remove — export the selection, or delete it and join what is left either side
• A crossfaded join — hide the seam where a removed section used to be
• Real zoom — the waveform is re-drawn from the audio at each zoom level, not stretched
• Zero-crossing snap — each edit lands where the waveform crosses zero, so the cut does not click
• Arrow-key nudging — 0.1 seconds a press, or 1 second with Shift held
• Fades — up to 5 seconds in and out, clamped to what the clip can actually hold
• Video files as input — cut the audio straight out of an MP4, MOV, MKV or WebM
• Output formats — MP3, M4A, AAC, OGG, FLAC and WAV, with bitrate, sample rate, channels and bit depth
• Batch with Pro — the same cut across a queue, delivered as one ZIP

Most online audio cutters decode your file, cut the samples and encode them again — even when nothing about the audio needed to change. That costs quality every time, because a lossy format re-encoded is a copy of a copy.

It is usually unnecessary. If the output format is the one the file already is, and no setting has been asked for that changes the samples, the encoded audio inside the selected span can be copied into a new file exactly as it is. Nothing is decoded, nothing is encoded, and the result is identical to the original — not close to it.

When the copy applies

• The output format matches the source: MP3 to MP3, FLAC to FLAC, and so on.
• The mode is Keep. Removing a section has to join two pieces, which needs decoding.
• No fade in, fade out or crossfade is set.
• The bitrate, sample rate, channel count and bit depth are all left as the source's.

What it costs

One thing: the edit lands on the nearest frame boundary in the original file rather than on the exact sample, because a frame is the smallest piece that can be copied without decoding. For MP3 that is about 26 milliseconds. Anywhere that matters, change a setting — or pick WAV or FLAC — and the cut is placed exactly instead. The tool says which of the two it is doing before you press anything.

Two different jobs share one set of handles. Keep exports the part you selected. Remove deletes it and joins the audio either side into a single continuous file — the cough in the middle of a take, the phone ringing during an interview, the sponsor read nobody asked for.

How the join is made

• Both pieces are timestamped from zero — otherwise the second piece keeps its original position and the gap it was meant to remove is written back in as silence.
• They are concatenated — by default the two pieces meet directly, which is right when the cut is placed in a quiet moment.
• Or crossfaded — a short overlap, typically 0.05 to 0.2 seconds, blends one into the other so nothing steps. The result is shorter than the two pieces by exactly the overlap.

Selecting the whole track in Remove mode leaves nothing behind, so the tool says so rather than handing back an empty file that looks valid.

A waveform drawn once and then stretched cannot show you anything you could not already see. Zooming here re-reads the audio over the visible span and draws it again, so at a two-second view individual syllables and drum hits are distinguishable, and an edit can be placed between them.

The controls that place an edit

• Drag anywhere on the waveform to start a new selection; drag within 12 pixels of an existing edge to move that edge instead.
• Arrow keys nudge the focused edge by 0.1 seconds, or by 1 second with Shift held. Home and End jump to the limits.
• Set start / Set end place an edge at the playhead, so you can listen for the moment and then mark it.
• Zoom to selection fills the view with the current selection plus a little margin, which is the fastest way to check both edges.
• Typed times accept a value to two decimal places and are bounded by the file's own length in both directions.

An edit placed at an arbitrary point lands part-way up a waveform. The signal jumps from wherever it was to zero in a single sample, and that step contains every frequency at once — which is what a click is. It is the most common complaint about cut audio and it has nothing to do with the format or the quality setting.

Editors have solved it the same way for decades: move the edit to the nearest point where the signal is already crossing zero, so the step has no height to begin with. Snapping is on by default, searches at most 10 milliseconds either way, and moves both edges in the same direction so the waveform stays continuous across a join. If there is no crossing nearby it leaves the edit exactly where you put it rather than dragging it somewhere arbitrary.

A fade solves the same problem differently, by ramping the level instead of finding a zero. A fade of half a second at each end is plenty for music; speech usually needs less. A fade longer than half the clip is reduced to fit and the tool says it did so, because a fade in and a fade out that overlap produce a shape neither control describes.

The fastest, highest-quality answer is almost always the format the file already is — that is the one that can be copied rather than re-encoded.

What each one is for

• MP3 — plays on everything. The right default for a clip you are going to send someone.
• M4A — AAC in an MP4 wrapper. Better than MP3 at the same bitrate, and native on Apple devices.
• AAC — the same audio as M4A but as a raw stream, for hardware that wants it that way.
• OGG — Vorbis, open and well supported on Linux and in games.
• FLAC — lossless and about half the size of WAV. The right choice for archiving a cut.
• WAV — uncompressed, universally readable, and what audio software expects to import. 16-bit free, 24-bit with Pro.

• Trim a podcast episode — remove the intro music and the sign-off without touching the audio in between.
• Cut a section out of an interview — delete a phone ringing or a long pause and crossfade the join.
• Make a highlight — pull a 30-second excerpt out of an hour-long recording.
• Extract audio from a clip — drop in an MP4 and cut its audio track directly.
• Prepare a sample — snap both edges to zero crossings and export as WAV for a sampler.
• Shorten a voice memo — cut the silence at the start and share a file a fraction of the size.

Free handles one file at a time, up to 50 MB and 30 minutes, and exports up to 44.1 kHz, 16-bit and 192 kbps. That is a real cutter, not a demo: every mode, every format and the lossless copy path are all available.

Pro removes the file count and the size cap — the ceiling becomes what your device can handle rather than a number we picked — and raises the export ceiling to 48 kHz, 24-bit and 320 kbps. A batch of cuts comes back as a single ZIP.

Every file is read, cut and written on your own device. Nothing is sent anywhere, which is why a 30-minute recording starts cutting immediately instead of after an upload, and why the tool works with no connection at all once the page has loaded.

Powered by optimised audio processing algorithms.

• MP3 Cutter
• WAV Cutter
• OGG Cutter
• M4A Cutter
• FLAC Cutter
• AAC Cutter

• Audio compressor
• Audio merger
• Audio splitter
• Audio converter
• Podcast trimmer
• Ringtone maker