MP3 cutter & audio trimmer

1. Drop your audio file. MP3, WAV, M4A, and OGG all work — anything your browser can decode. The file is read locally and its waveform is drawn so you can see exactly where the quiet and loud parts are.
2. Mark the selection. Drag the Start and End sliders, or type precise times like 1:23.4 into the minute:second boxes. The selected region lights up on the waveform, and the selection badge shows its exact length.
3. Preview it. The preview button plays only the selection — including any fades — so you hear precisely what the exported file will contain before you commit. Click again to stop.
4. Export and download. Choose MP3 (128 or 192 kbps) or lossless WAV and click export. The cut is rendered and encoded on your device and downloads as yourfile-cut.mp3.

Everything stays adjustable after loading the file — move the markers, toggle fades, switch formats — and re-export as many different cuts as you like from the same source.

The classic use for an audio trimmer is the ringtone: most phones want a clip of roughly 20–30 seconds, and the part you actually want is almost never the first 30 seconds of the track — it's the chorus or the hook somewhere in the middle. The waveform makes that easy to find (choruses are visibly louder), the mm:ss.d inputs let you land the cut on the exact beat, and a half-second fade in keeps the ringtone from starting with a jarring click. Export at 128 kbps for a small file, or 192 kbps if you want it indistinguishable from the original on phone speakers.

For iPhone users, one honest practical note: iOS expects ringtones in M4R format and they must be added through Finder/iTunes or GarageBand — so export your cut here as MP3, then let GarageBand do the final conversion step. Android phones accept MP3 ringtones directly: export, copy to the phone, and pick it in Settings → Sound. Either way, the trimming — the part that actually requires judgment and precision — happens here, for free, without uploading your music anywhere.

Ringtones aside, most real trimming jobs are about cleaning up speech. A phone voice memo starts with five seconds of fumbling and ends with a shuffle as you reach to stop the recording; an interview needs the pre-chat sliced off before you send it to a client; a lecture recording has one crucial two-minute answer buried at minute 43. With this tool you find the moment on the waveform (speech shows as dense bars, silence as flat gaps), set the markers, preview to verify, and export just that segment. Because the export is a fresh render of the selection, the output contains nothing but the part you chose — there's no hidden metadata or leftover audio from the rest of the file.

The fade in / fade out options (0.5 to 2 seconds each) matter more for these cuts than people expect: cutting speech or music mid-waveform produces an audible click or an abrupt stop, and even a half-second fade smooths that into something professional. For clips headed to social media, a one-second fade out is the difference between "clip ends" and "clip was produced."

MP3 is the right default for anything that gets shared or set as a ringtone: it's universally supported and small. At 128 kbps a 30-second clip is roughly 470 KB — ideal for messaging apps and ringtones. At 192 kbps the same clip is about 700 KB with noticeably better treble on good headphones — the safer pick for music you care about. The MP3 encoding happens entirely in your browser using lamejs, a JavaScript port of the LAME encoder, which is why no server is involved even for this step.

WAV is uncompressed PCM — bit-for-bit lossless, and much larger (about 5 MB per 30 seconds of stereo at 44.1 kHz). Choose WAV when the clip is headed into further editing: an audio editor, a video project, a DAW, or a podcast production chain. Re-encoding an MP3 to MP3 again at each step degrades quality each time; cutting to WAV keeps the audio pristine until the final export. One thing this tool won't pretend to do: it can't restore quality your source never had — cutting a 128 kbps MP3 and exporting at 192 kbps produces a bigger file, not a better-sounding one.

The waveform is drawn by decoding the entire file with the Web Audio API and plotting the average loudness of each slice of time as a vertical bar. Reading it takes about ten seconds to learn: tall dense sections are loud passages — choruses, applause, raised voices; short bars are quiet passages; flat gaps are silence — the pauses between sentences, the gap before the song starts. That visual map is what makes precise cutting possible without scrubbing back and forth blindly. The highlighted band shows your current selection, and the edge markers move live as you drag the sliders, so lining a cut up against the start of a phrase or the last beat of a bar is a matter of nudging until the marker sits in the silence gap next to it.

• Ringtone: 20–30 seconds starting at the chorus. Fade in 0.5s so it doesn't start with a click, no fade out (the phone stops it anyway), MP3 at 128 kbps.
• Podcast promo clip: 30–60 seconds around the best quote. Fade in 0.5s, fade out 1s so the clip ends gracefully mid-conversation, MP3 at 192 kbps for clean voice reproduction.
• Alarm or notification sound: 5–15 seconds of something that starts strong. No fade in — an alarm should hit immediately — short 0.5s fade out, MP3 at 128 kbps.
• Sample for further editing: cut generously (a second or two of margin on each side), no fades — you'll shape the ends in the editor — and export WAV so no quality is lost before the real work starts.

In every case the preview button is your friend: it plays exactly what will be exported, fades included, so there are no surprises after downloading.

Under the hood, the export doesn't splice compressed MP3 frames — it re-renders your selection through an OfflineAudioContext, the browser's sample-accurate offline audio engine. The selection's start and end are converted to exact sample positions (at 44.1 kHz, a tenth of a second is 4,410 samples), the fades are applied as gain ramps in that render, and only then is the result encoded to MP3 or WAV. That architecture is why the mm:ss.d inputs can be trusted literally, why fades sound smooth rather than stepped, and why the exported duration matches the selection badge to the decimal. It's the same processing model desktop audio editors use — just running in a tab.

Audio files are personal: voice memos, band demos that aren't released, client interviews under NDA, therapy notes, family recordings. Most online audio cutters upload your file to a server to process it — check their privacy policies for how long the copy is retained. This tool is built the other way: the file is read with the File API, decoded by your browser's own audio engine, trimmed in an OfflineAudioContext, and encoded with JavaScript — all inside this tab. There is no upload endpoint, no progress bar labeled "uploading," and no server copy to worry about, which you can verify by loading the page and then going offline: everything still works.

The honest trade-off is that your device does the work. Decoding loads the full uncompressed audio into memory — roughly 10 MB per minute of stereo audio — so a three-minute song is trivial, an hour-long podcast is fine on a laptop, but a four-hour recording may be slow or fail on a low-memory phone. For files that size, a desktop editor like Audacity is the better tool, and we'd rather tell you that than let the tab struggle.

• Anyone making a ringtone — cut the chorus, fade it in, export MP3, done.
• Podcasters pulling a quotable 30-second clip out of an episode for social promotion.
• Journalists and researchers extracting the relevant answer from a long interview recording before filing or transcribing it.
• Musicians trimming rehearsal recordings and demos to the take worth keeping, exporting WAV for the DAW.
• Content creators cutting intro music, voiceover segments, and sound effects to length for videos.
• Teachers and students clipping lecture recordings down to the section that's actually on the exam.

If your clip is destined for a web page alongside images and documents, the image compressor and file converter cover the rest of the media pipeline the same private way.