Make music in your browser: an online studio with no download

What Gusta Music is

Most music software wants a download, a licence and an evening of setup before it makes a sound. Gusta Music wants a browser. It is a small digital audio workstation for people who have an idea and forty minutes: sketch a beat, put chords under it, record a vocal, mix it, export a file. Underneath, a song is a plain document, and every change, yours or the AI's, goes through one editing layer, which is why undo always works.

Who it is for

Hobbyists who want to make beats in the browser without committing to a big install. Songwriters who want to hear an idea arranged rather than described. Anyone on an iPad who has been told real music software does not run there. It is where an idea becomes a song instead of a voice memo.

What is inside

A timeline that behaves like a timeline

Tracks stack down the left with volume, pan, mute, solo, a record-arm dot and a live meter. Clips drag and resize across the ruler, and dragging on the ruler sets a loop region so you can work four bars until they are right. Section markers label the arrangement, and the tempo can change across the song rather than being one number.

Sounds you pick rather than build

Instead of a wall of synthesis parameters there is a sound browser on every track head: families of designed sounds with variations, and tapping one plays it through that track's real effect chain, so you hear exactly what you are choosing. Build a sound you like and you can save the whole track voice under a name, then reuse it in any project or ask the copilot for it by name.

Patterns, not note-by-note drudgery

Drum patterns come from a library organised by genre, with an intensity control that thins or thickens the layers and an option to drop a fill every few bars. Chords are entered as symbols and voiced for you. The strum editor works the way a guitarist thinks: chords per bar on top, a grid of down and up strokes underneath, rendered into real guitar voicings. Melodies can be written as scale degrees, which keeps them in key by construction.

Recording, with a microphone or a keyboard

Arm a track and hit record: you get a count-in, a live input meter, optional monitoring and a ghost clip that grows as you sing, and the take lands in your sample library and on the timeline at once. With a MIDI keyboard, notes play the armed track's instrument live and are captured as a note clip rather than audio.

Your own audio, understood

Import an MP3 or a WAV and it is analysed: samples over eight seconds get a detected tempo and key, so dropping one in stretches it to your song's tempo instead of fighting it. Stem tools split a track into drums and music, reduce a centred vocal, or tune a take to the key of the song, all computed in your browser rather than uploaded anywhere. A one-click vocal pass tunes the lead in place and lays doubles and a harmony underneath, as one undo step.

Mixing, and getting the song out

Every track runs the same chain: distortion, a six-band EQ with a drawable curve, filter, chorus, delay, reverb, then volume, pan, mute and solo into the master bus, plus automation lanes. Export uses those same chain builders, so the file you download is the mix you heard: a master as 16-bit or 24-bit WAV or MP3, every track as its own 24-bit WAV in a zip, a standard MIDI file with drums on channel 10, or the project file with its samples bundled in.

What makes it different

The AI uses the tools instead of talking about them. Gusta does not hand you a paragraph about what a lo-fi groove should contain. It dispatches the same edits your clicks do: adding a track, choosing a sound, writing a drum pattern, voicing chords. Each one lands on the timeline as an ordinary change you can select, nudge or undo, and the clips it touched are marked so you can see which parts came from the machine.

Your work stays yours. This is not prompt-to-song. No model hands you a finished audio file you are then licensed to use. There is a document full of notes, samples and settings, and you are its author. That distinction is the point.

Local first. Projects live in your browser by default and samples are stored on your device. Audio analysis, stem separation and pitch correction all run locally. Signing in adds cloud saves and the copilot; it does not move your work somewhere you cannot reach it.

Free to start. The studio is free. Only AI work spends tokens, and during the open beta those tokens are granted rather than sold.

Free tools, no account needed

Some of what the studio does is useful on its own, so those parts have pages of their own. Each one runs entirely in your browser: the audio never leaves your device, there is nothing to install, and you do not need an account.

• BPM detector Drop in a track and read its tempo, from the same onset analysis the studio uses when you import a sample.
• Song key finder Find the key and the chords that fit it, using a chromagram and the Krumhansl profiles rather than a guess.
• Vocal remover Cancel the centre channel to pull a centred vocal out of a stereo mix, with the bass kept rather than hollowed out.
• Stem splitter Separate the percussive hits from the sustained parts and download both as WAVs.
• Metronome An accurate click with adjustable tempo, time signature and accent, scheduled on the audio clock.

Questions people ask

Also from Gustalabs: Gusta 3D, parametric CAD for 3D printing, and Gusta Draw, an online vector and SVG editor. One account across all three.