Guide · 8 min read · free tool
Trim Audio in the Browser: Keep the Take, Cut the Chair Squeak
Sep 4, 2026 · Zelpura Editorial · All guides
Most audio edits a solo creator needs are not mixes. They are a start point and an end point: the breath before the sentence, the 'are we recording' at the top, the chair at the end. Sending that file to a website so it can trim four seconds is a bad trade. This page decodes the audio in your browser, keeps the range you set, and encodes a download. The take does not become an account. You still have to listen to the cut, because a trim that clips the first consonant is worse than the chair squeak you were trying to remove.
The inputs are the job, not a formality. You pick an audio file the browser can decode, then a start and an end. Start is where the useful sound begins, not where the waveform first twitches if that twitch is silence you want to keep for a breath. End is the last moment you want a listener to hear. If you are trimming a music bed under a voice, this page is the wrong tool. It keeps one range of the whole file. It does not separate stems, remove a hum, or make you a better speaker. Those are different jobs. This job is the cut.
A voice memo that starts with two seconds of fumbling and ends with a door. Set the start just before the first word, the end just after the last, and leave a tenth of a second of air so the edit does not feel like a guillotine. Play the result. If the first syllable is missing, move the start earlier. Ears are better than waveforms for speech. Waveforms are better for finding the long silence. Use both. Then name the download with the episode and the take, not trimmed.mp3. Future you will have twelve trimmed.mp3 files and no idea which one was the good cold open.
The miss is trimming the only copy. Export is a new file. Keep the original until the episode is published, and maybe longer. The second miss is cutting so tight that a loudness processor later makes the edit click. Leave a little air. The third miss is uploading a private conversation to a 'free trimmer' because this page failed on an exotic format. If a file will not decode, convert it locally with a tool you installed, or accept that this browser cannot read it. Do not solve a codec problem by giving the audio to a stranger. The content is often more sensitive than the user remembers. Voice memos are diaries with a file extension.
Duplicate the file or trim from a copy. Set in and out. Listen once with eyes closed. If you wince, move the point. Download, then listen to the download, not only the preview. Previews lie in small ways. The file you will upload to a host is the download. For a podcast, trim the raw, then do loudness and music in the editor you already trust. Do not ask a trimmer to be a studio. A studio that runs in a tab and uploads your interviews is a studio you should not trust with other people's voices. Guests did not consent to a random processor.
This is a local cut, not a restoration suite. It will not remove noise, fix clipping, or separate a song you do not have the right to sample. If you do not own the audio or have permission, do not trim it into a new distribution. A shorter infringement is still an infringement. Some formats will not decode in every browser. That is a limit, not a prompt to find an upload site. Nothing here stores your file on a server of ours. Your browser's own memory is still your machine. Close the tab when you are done if the audio was private.
Keep a naming scheme: episode-raw, episode-trim, episode-master. The trimmer belongs in the middle of that chain, not as the place files go to get mysterious new names. When a guest asks you to cut a sentence, do it from the raw and note the timestamps. A note is how you show the cut happened. A mystery file is how you wonder, six months later, whether you ever did. The tool is fast. The archive is the part that makes the speed useful.
Read the result back before you act. Play the first second and the last second of the download. Those are the only seconds a trimmer can ruin. If they are clean, send it. If they are clipped, move the points and do not talk yourself into 'good enough' on a cold open.
A video's audio and a standalone recording are different files. Trimming the audio does not trim the video. If you need both, use the video trimmer on the video and keep this page for the voice memo and the podcast stem. Mixing those jobs is how the picture and the sound end up different lengths. Different lengths are a publish-day surprise. Surprises on publish day are almost never good.
Related: Video trimmer · Podcast payout · PDF tools · All 93 tools