CHAIN TOOLS INTO A RECIPE
Audio recipes
Run several edits in a row on one recording, or a whole folder of them, and download one finished file. Save the steps as a link and reuse them on next week’s episode.
Hear a voice with hiss and a faint hum, then run the recipe on it. Music recipes use a short original beat instead.
Podcast episode cleanup
- Reduce noise
Spectral gating against the quietest tenth of the recording: hiss, fans, hum and room tone.
- Remove silence
Shortens every pause longer than the minimum, keeping a little padding around speech.
- Normalize loudness
One gain to reach an integrated loudness (LUFS, ITU-R BS.1770) or a peak level, with a soft limiter at the ceiling.
- Fade in and out
Smooth (sine) or straight fades at the start and the end.
Export, once at the end
Ready-made recipes
Each one opens in the editor above. Change anything, then run it on your own file.
- Podcast episode cleanupLower hiss and rumble, shorten long pauses, level to −16 LUFS and finish with short fades. A 128 kbps MP3, the size most podcast hosts suggest for speech.Reduce noise → Remove silence → Normalize loudness → Fade in and out → MP3 128 kbps
- Voice memo to clean MP3Cut the tap at each end, drop dead air, mix to mono and level for a phone speaker. Small enough to email.Trim start and end → Reduce noise → Remove silence → Mix to mono → Normalize loudness → MP3 96 kbps
- Lecture recording for reviewReduce room noise and hum, cut pauses longer than 1.5 s, play 1.25× faster without the chipmunk voice, and keep the file small.Reduce noise → Remove silence → Change speed → Mix to mono → Normalize loudness → MP3 64 kbps
- Audiobook chapter, ACX-style levelNoise reduction, then an average level of −20 dB RMS with peaks held below −3 dB, exported as a 192 kbps, 44.1 kHz MP3. Pauses and room tone are left alone.Reduce noise → Change volume → MP3 192 kbps
- Song to slowed + reverbSlow to 0.85× with a deep room, bring it to streaming loudness and let it fade out over three seconds.Slowed / sped up + reverb → Normalize loudness → Fade in and out → MP3 192 kbps
Run a recipe in three steps
- Pick a ready-made recipe below or build your own: add steps, change their settings, and use the arrows to put them in order.
- Choose one or more audio or video files, or try the sample. Press Run recipe and keep this tab open while each step runs.
- Compare the before and after waveforms, levels and players, then download the file (or a ZIP for several). Copy the link to use the same recipe again.
What a recipe can do
Each step is the same processing that runs on its own tool page: trim a fixed amount from the start and end, remove silence, reduce steady noise, normalize loudness to a LUFS or peak target, set a fixed gain or an average RMS level, mix to mono, change speed without the chipmunk voice, change pitch, slow down with reverb, boost bass, add fades, or reverse. Steps run in the order you list them, and order matters. Reducing noise before removing silence lets the silence detector see real pauses instead of hiss. Normalizing near the end means the level you hear is the level you get, after the speed change or fade has done its work.
Encoded once, at the end
Your file is decoded to 32-bit floating-point samples once. Every step works on those samples, and only after the last step is the result encoded to MP3, WAV, M4A or OGG at the bitrate you choose. Doing the same job by downloading an MP3 from one tool and loading it into the next would re-encode the audio at every hop and stack up compression artefacts, the swirly highs and smeared consonants you hear on a file that has been through a converter too many times. If you want no lossy step at all, choose WAV.
Sharing a recipe
The editor keeps the recipe in the page address after #r=. It is a short code that holds a version number, the step names and any setting you changed from its default. Copy the link and send it to a co-host, or bookmark it for yourself: opening it loads the same steps, ready to run on whatever file you choose. The part after # is never sent to our server by the browser, and it contains no audio and no file name.
About the ACX audiobook recipe
ACX, the Audible platform for audiobook producers, publishes its audio submission requirements: an RMS level between −23 and −18 dB, peaks no higher than −3 dB, a noise floor no higher than −60 dB, a 192 kbps or higher constant bitrate MP3 at 44.1 kHz, and at most 120 minutes per file, with room tone at the start and end. Our recipe reduces noise, sets the average level to −20 dB RMS with peaks held under −3.5 dB, and exports a 192 kbps, 44.1 kHz MP3. After it runs, the page compares the result with those numbers. It does not remove pauses, so your room tone stays. It cannot make a noisy room meet the noise-floor limit, our noise reading is an estimate, and ACX’s own review decides whether a file passes.
Limits
All steps run in your browser’s memory. A one-hour stereo recording needs over a gigabyte of working memory once decoded, so long files are better on a computer than a phone. Many files are processed one at a time with the same recipe. The trim step cuts fixed amounts from each end, so it suits recordings with the same lead-in every time; to choose a section by hand, use the audio trimmer. Video files are accepted, but a recipe exports audio only.
Frequently asked questions
Why not just run each tool one after another?
You can, but each download is a fresh lossy encode. A recipe keeps the audio uncompressed between steps and encodes once, so quality holds up better, and it saves loading the file into five pages.
What does a recipe link contain?
A version number, the list of steps and the settings you changed, packed into a short code after #r= in the address. No audio, no file name and nothing about your device. Opening the link loads the steps; you still choose the file.
Can I run one recipe on many files at once?
Yes. Choose several files and press Run recipe. They are processed one after another with identical settings, and you can download each result or all of them as a ZIP.
Does the audiobook recipe guarantee ACX approval?
No. It levels the audio to ACX’s published RMS and peak numbers and exports the file format they ask for, then shows how the result measures against their list. Background noise, room tone and the human review are up to your recording.
Will an old recipe link still work after the site changes?
Each link carries a version number. If we ever change how steps or their defaults work, older links are upgraded when they load, and a link from a newer version says so instead of running the wrong settings.
All processing runs locally in your browser memory via WebAssembly and Web Audio. Files never leave your device. Built-in peak limiting prevents audio clipping.
AudioClover Engineering · Updated October 2026KEEP IN THE LOOP
A better workflow for your next recording.
Get tool updates and audio tips. Batch presets and a podcaster plan are on the roadmap.