Booth Free voice recorder & AI noise removal

On this device FAQ nader.io
Ready 48 kHz · Mono

0:00.0

R

Takes

Record, or drop an audio file here.

    FAQ

    Why Booth?

    I created Booth because I was unhappy with the Austrian dialect that was generated for my game Diebstahl Simulator. I actually wanted to record my own, but I didn’t have audio software, and I wanted applying effects to be easy instead of digging through super complicated audio software. So I launched this. I hope you like it.

    Is Booth free?

    Yes. No account, no sign-up, no watermark and no usage limits.

    Are my recordings uploaded?

    No. Recording, noise removal and mixing all run in your browser, and takes are stored on your device. Nothing is sent to a server. Clearing the site data deletes them.

    How does the AI noise removal work?

    Booth runs DeepFilterNet3, an open-source neural network trained to separate speech from noise, on your device. It removes hum, fans, traffic, keyboard and room tone while keeping the voice at full 48 kHz quality. The model downloads once, the first time you enhance.

    Can I clean up a recording I already have?

    Yes. Drop an audio file anywhere on the page or use Import. WAV, MP3, M4A, OGG, FLAC and WebM work in most browsers.

    Does Booth cut the dead air at the start and end?

    Yes. Enhance finds where you start and stop talking and cuts everything before and after, including a mic bump, a click or the keyboard while you reach to stop recording. Pauses in the middle are kept, and the original take stays untouched.

    Why do all takes come out at the same volume?

    Every enhanced take is measured with ITU-R BS.1770 and set to −16 LUFS integrated loudness, the level Apple recommends for podcasts, with peaks held under −1 dB. Different takes, microphones and rooms end up equally loud.

    What do the presets do?

    Natural, Podcast, Radio, Warm and Room set noise removal, EQ, de-essing, compression and room together. Supermarket turns your voice into a store announcement, the way the tannoy sounds in Diebstahl Simulator. Fine-tune changes each one, and the take you are working on updates as you move a slider.

    Can I play several takes at once?

    Yes. Turn on Layer above the takes and every take you start plays on top of the others. Turn on Loop and they go round and round without a gap, so a few short takes become a beat or a backing track. Stop all ends them together.

    Can loops play in time with each other?

    Yes. Turn on Sync next to Layer and Loop. The first take you start sets the tempo, or type a BPM or tap it in. Every loop is then cut or padded to a whole number of beats or bars, so loops never drift apart, and a take you start while others play waits for the next bar. Click adds a metronome.

    Can I add an echo?

    Yes. Fine-tune has Echo for how loud it is, Echo time for the gap between repeats and Feedback for how many there are. It is part of the enhanced WAV you download. With Sync on, Echo time snaps to note lengths at the tempo, such as 1/8 or a dotted 1/8, and on a loop the echo rings on into the next turn instead of being cut off.

    Can I send someone a link with my sound?

    Yes. Picking a preset puts it in the address, for example nader.io/projects/booth/?sound=supermarket. Whoever opens that link starts with the same preset selected, ready to record.

    Are there keyboard shortcuts?

    R starts and stops recording. The up and down arrows move between takes, and Space plays or pauses the selected one without scrolling the page. L turns Layer on and off, O does the same for Loop and S for Sync, T taps the tempo, and Escape stops everything. Double-click a name to rename a take.

    What format do I get?

    Enhanced takes download as 16-bit, 48 kHz WAV, ready for any editor or podcast host. Originals download in the format they were recorded or imported in.

    Delete all takes?

    Every take on this device, original and enhanced, is deleted. This cannot be undone.