Audacity is free and it is what most podcasters and small stations actually edit in. What it does not ship is a loudness meter that reads LUFS, a broadcast voice chain, a de-clicker or a caller repair. Airchain adds all nine, free, in about a minute.
No account, no email, no watermark. GPL v2+ — the same licence as Audacity itself.
Eight effects and one analyser, built for spoken word rather than adapted from music tools. Tap any card to see its settings.
Measure before you touch anything
Hit the number Spotify and your regulator want
The broadcast chain, in one effect
Four noises, four right tools
Mouth clicks and lip smacks, not your consonants
Make a phone-in sit next to the host
The dry run Truncate Silence never had
Duck the bed the way imaging is actually cut
Seven voice treatments imaging is built from
Tells you where you stand. Integrated loudness measured the way your delivery platform measures it — gated, to ITU-R BS.1770 and EBU R 128 — plus short-term maximum, loudness range, peak level, headroom, and a plain sentence saying how far off target you are.
In practice: Run this first on a finished episode. Every other plugin in the pack is easier to set once you know what the recording is actually doing.
Every control, exactly as it appears in Audacity.
Lands the selection on a delivery target in one move: −16 LUFS for podcast, −19 for a mono master, −14 for music streaming, −23 for EBU R 128 broadcast, −30 for a music bed under a voice. Then holds peaks under a true peak ceiling with a transparent brick-wall limiter.
In practice: The never add more than control is the one that saves you: it stops a quiet recording being boosted so far that the room comes up with the voice.
Every control, exactly as it appears in Audacity.
Low cut, de-esser, compressor, voicing EQ, make-up gain and a peak ceiling — in the order a radio processing chain uses them. Five voices to pick from, then a single Strength dial does the rest. Prefer to dial it in by hand? The same numbers are in the Audacity compressor settings guide.
In practice: Every control is absolute. What the dialog shows is what is applied, so there are no hidden preset offsets to reason about.
Every control, exactly as it appears in Audacity.
Rumble, mains hum, hiss and room tone, each handled with the correct tool rather than one broadband noise reduction that dulls the voice. Hum gets a narrow notch on 50 or 60 Hz and its harmonics. Room tone gets an expander that pushes the air down between words.
In practice: Everything ships off or mild. Turn up only what you can actually hear — and if you are not sure what you are listening to, measure a silent pause with Airchain Meter first.
Every control, exactly as it appears in Audacity.
Ducks only the band the mouth noise lives in, only while it lasts. The body of the voice is untouched, so a read does not go dull the way broadband noise reduction makes it go dull. Detection is relative to the recording's own high-frequency level, so one setting holds across a loud read and a quiet one.
In practice: Three modes by speed: mouth clicks, the slower lip smacks, and the fastest setting for vinyl crackle and ticks.
Every control, exactly as it appears in Audacity.
A caller arrives band-limited, codec-smeared, too quiet and the wrong colour. This clears the rumble and the codec fizz, rebuilds the presence the band limiting took away, evens the caller out, and lands them on the same loudness target as the rest of the programme.
In practice: Run it on the caller's track only, with the host muted or on another track. Four source types: landline, mobile, VoIP and laptop mic.
Every control, exactly as it appears in Audacity.
Report only measures every pause and tells you how much time you would recover, then prints the exact numbers to type into Audacity's own Truncate Silence. Nothing is changed. Clean the gaps takes the room tone out of the pauses without touching a word and without changing the length.
In practice: Breaths at the edge of each pause are protected and every edge is ramped, so cleaning never clips the front of a sentence.
Every control, exactly as it appears in Audacity.
Audacity's Auto Duck needs the voice on the track below and ducks the whole timeline. This works the other way round: select the stretch of bed that sits under the voice and that stretch is ducked, with the ramps at each end that make a duck sound deliberate rather than accidental.
In practice: Voice space lifts the bed out of the band the voice occupies. It buys more intelligibility than another 3 dB of duck ever will.
Every control, exactly as it appears in Audacity.
Telephone, mobile, walkie-talkie, AM radio, megaphone, intercom and vintage set. Each one is a band limit, a saturation and a noise bed chosen to sound like the thing it is named after, rather than a generic band-pass with the mids scooped out.
In practice: Keep Mix at 80–85% on a promo: enough effect to read as a phone call, enough of the dry voice to survive a car radio.
Every control, exactly as it appears in Audacity.
Nyquist plugins are plain text files. Nothing is compiled, nothing asks for admin rights, and nothing phones home.
Windows: double-click the .bat inside install. Mac: right-click the .command and choose Open, then Open anyway — macOS asks because the file came from the internet.
Audacity only scans for new plugins at startup, so quit it completely and reopen.
Audacity 4: Settings → Plugins. Audacity 3: Tools → Plugin Manager. Tick the nine Airchain entries. New Nyquist plugins always arrive disabled, so this step is not optional.
Prefer to do it by hand? Drop the nine .ny files into Audacity's plug-in folder, or use Tools → Nyquist Plugin Installer from inside Audacity. Full paths for every platform are in the INSTALL.md inside the download.
Order matters more than settings. Loudness goes last, because everything before it changes the level.
Yes. All nine plugins are free to download and use, with no account, no trial and no watermark. They are released under the GNU General Public License v2 or later, which is the same licence Audacity itself uses, so you also get the source and the right to pass them on.
Audacity 3.2 and every version since, including Audacity 4. They are Nyquist plugins, so the same files work on Windows, macOS and Linux with nothing to compile and nothing to authorise.
Download the pack for your system and run the installer inside it, which copies the nine .ny files into Audacity's plug-in folder. Then restart Audacity, open the plugin manager (Settings > Plugins in Audacity 4, Tools > Plugin Manager in Audacity 3) and enable the nine Airchain entries. Audacity always discovers new Nyquist plugins as disabled, so that last step is not optional.
Eight of them appear under the Effect menu. Airchain Meter appears under Analyze, because it measures rather than changes. If the menu entries look greyed out, you have no project with a track open — Audacity disables effect commands on an empty project.
−16 LUFS integrated for a stereo podcast, −19 LUFS for a mono master, with a true peak ceiling of −1 dBTP. Broadcast radio in Europe uses EBU R 128, which is −23 LUFS. Airchain Meter reads all of these and Airchain Loudness hits them in one move.
Not on its own. Audacity ships noise reduction, which is the wrong tool for clicks and tends to dull the voice. Airchain De-Click ducks only the band the click lives in and only while it lasts, and Airchain Voice has a level-independent de-esser built into the chain.
They solve the other half of the job. AI JingleMaker makes the imaging — jingles, sweepers, station IDs, podcast intros. These plugins clean, level and finish the voice tracks you cut around that imaging in Audacity.
No. They exist because the people who use AI JingleMaker are mostly editing in Audacity, and a cleaner finished episode makes the imaging sound better too. Take them and use them even if you never buy a jingle.
Turn your idea into a finished jingle, radio ID, or promo. Make one yourself in seconds with our AI Jingle Maker — or have a custom one handcrafted for you.