Edit audio by editing text
This free podcast and audio editor turns your recording into text, word by word. Delete a word from the text and it’s cut from the sound. Select a sentence you don’t want, press Delete, and it’s gone, with a smooth join so nobody hears the cut. It’s made for podcasts, interviews, voice-overs, lectures, audiobooks, voice memos and the sound of any video.
One click removes every “um” and “uh”, shortens long pauses, and removes background noise like hiss, hum, fans and traffic. Then set podcast loudness, add intro and outro music, and download an MP3, WAV or M4A along with the transcript and subtitles.
Everything runs in your browser. Your recording is never uploaded, and there’s no account, no watermark and no limit on how many files you edit.
In the editor, How to use in the top bar opens a short guide, and the search box finds any option: type what you want to do, like “remove um” or “make it louder”, and it takes you there.
How to edit audio by editing text
- Upload or record. Drop an audio or video file on the page (MP3, WAV, M4A, OGG, FLAC, MP4, MOV or WebM, up to 4 hours), or click Record with your microphone.
- Wait a moment for the transcript. The words appear while the rest is still being transcribed, so you can start editing right away.
- Cut by deleting text. Select words and press Delete, or click Cut in the bar that appears. Cut words stay in the text, crossed out, so you can click one and choose Restore to bring it back.
- Clean up in one click. In the Clean up tab, remove filler words, shorten long pauses, remove repeated words and remove background noise.
- Make it sound good. In the Sound tab, set the loudness for podcast apps, even out loud and quiet voices, add clarity, fades and intro or outro music.
- Export. Click Export and download MP3, WAV or M4A, plus the transcript (TXT) and subtitles (SRT or VTT) timed to the edited audio.
Click any word to play from there. Press O or click Original to hear the recording before your edits, and press Ctrl+Z to undo anything.
One-click clean up
- Remove “um” and “uh”: finds filler words in the text, and also short “um” sounds the speech model didn’t write down, marked as um?. Remove them all at once, then click any one and choose Restore to keep it.
- Remove “like”, “you know” and “I mean”: finds filler phrases only where they’re set off by commas, so “I like it” stays. Look at the matches first, then cut them all or one by one.
- Shorten long pauses: choose how long a pause has to be (for example over 1 second) and how long to make it (for example 0.5 seconds). Pauses show in the text as small gray times; click one to shorten just that pause.
- Remove repeated words: finds stutters like “I I think” and “the the” and cuts the first copy.
- Remove background noise: a small AI model removes hiss, hum, fans, traffic and room echo while keeping the voice. Set the strength, and compare with the original in one click.
- Trim the start and end: cuts the silence or noise before the first word and after the last one.
Make it sound like a podcast
- Loudness: Podcast sets −16 LUFS, what podcast apps expect. Quiet (−19 LUFS) suits audiobooks and Loud (−14 LUFS) suits video sites. Peaks are kept below −1 dB so nothing distorts.
- Even out volume: brings quiet and loud voices closer together, so a guest far from the microphone is as easy to hear as the host.
- Voice clarity: makes voices clearer and less muddy.
- Remove low rumble: removes wind, desk bumps and air-conditioning rumble.
- Fades: fade in at the start and fade out at the end.
- Intro and outro music: add a music file that plays before you start talking and fades down under your voice, and one that rises as you finish.
What you hear in the editor is exactly what you download: cuts, noise removal, loudness, fades and music are all included in the export.
Fix the transcript, then download it
- Fix text: select a misheard word or name and press E to correct it. The sound stays the same.
- Find: press Ctrl+F to find a word or phrase, then cut or restore every match at once.
- Hide cut words: see the finished text without the crossed-out words.
- Transcript and subtitles: download the text as TXT, with or without times, or as SRT and VTT subtitles. The times match the edited audio, so the subtitles line up with your finished file.
Languages
Transcription works in English, Spanish, French, German, Italian, Portuguese, Dutch, Hindi, Arabic, Russian, Ukrainian, Polish, Turkish, Swedish, Japanese, Korean, Chinese, Indonesian, Vietnamese and Urdu. Choose the language and the accuracy before you upload, or change them later with Transcribe again.
The speech model is downloaded once and kept by your browser: Fast is about 40 MB, Balanced about 80 MB and Most accurate about 250 MB.
Your recordings are private and saved
Transcription, editing and export all happen on your own device, so your recordings never leave it. Your projects are saved in this browser as you work, under Your recordings on the start screen, so you can close the tab and come back later to keep editing in the same browser.
Tips for better recordings
- Speak close to the microphone and record in a small room with soft furnishings, so there’s less echo to remove.
- Leave a short pause after a mistake, then say the sentence again. The pause makes the mistake easy to find and cut.
- Clean up first, then listen once. Remove fillers and long pauses, then play the whole thing and restore anything that sounds rushed.
- Keep a breath of space. Shortening pauses to about half a second keeps the talk natural. Shorter can sound rushed.
- Set the loudness last, after the cuts and noise removal, so the level is measured on the finished sound.
Frequently asked questions
Is this audio editor free?
Yes. It’s free, with no sign-up, no watermark and no limit on how many recordings you edit or download.
How does editing audio by editing text work?
Your recording is transcribed word by word, and each word knows where it is in the sound. When you delete words from the text, the editor cuts that part of the sound and joins the rest with a short crossfade, so the cut is smooth.
Are my recordings uploaded?
No. Transcription, editing, noise removal and export all run in your browser, and your recordings stay on your device. Only the tool’s own files are downloaded, such as the speech model and the noise removal model.
Can it remove “um” and “uh” automatically?
Yes. Open the Clean up tab and click Remove fillers. Every “um” and “uh” is cut at once, and you can click any one in the text and choose Restore to bring it back.
Can it remove silence from audio?
Yes. Under Long pauses, choose how long a pause has to be and how long to make it, then click Shorten. You can also click any pause in the text to shorten just that one.
Can it remove background noise?
Yes. Turn on Remove background noise in the Clean up tab. A small AI model removes hiss, hum, fans, traffic and echo while keeping the voice, and you can set the strength.
Can I edit the sound of a video?
Yes. Upload an MP4, MOV or WebM video and the editor opens its sound. Edit it like any recording and export the sound as MP3, WAV or M4A.
Which file formats can I export?
MP3 (128, 192 or 320 kbps), WAV and M4A for the sound, in mono or stereo. The transcript downloads as TXT, and subtitles as SRT or VTT.
How long can a recording be?
Up to 4 hours. Longer recordings take more time to transcribe; the words appear as each part is done, so you can start editing early.
Which browsers work?
A recent version of Chrome, Edge, Firefox or Safari on a computer works best. It also works on phones and tablets, though transcribing a long recording is slower there.
