AI stem separation
Vocals, drums, bass, guitar, piano — Demucs splits any song locally on your machine. No cloud, no upload.
FOR WINDOWS · FREE · OPEN SOURCE
Audiola splits any song into its tracks with AI, lets you arrange, sing along, swap voices and master — in a studio that actually feels like one.
Vocals, drums, bass, guitar, piano — Demucs splits any song locally on your machine. No cloud, no upload.
Cut, drag, fade and loop clips. Undo history, snapping, range effects — a real workspace.
Karaoke view with a note band and scoring. "Make it perfect" gently pulls your take onto the original melody.
EQ, compressor and target loudness per EBU R128 — with live A/B and 14 profiles from streaming to club.
Clone your own voices (locally or via ElevenLabs), text-to-speech tracks, per-clip voice swap.
Place tracks freely in space — an immersive 7.1.4 bed for headphone moments you won't forget.
THE STUDIO
Drop an MP3 and it becomes a track. Split it into stems, cut clips at the playhead, move parts between tracks, set fades with a single grab.
.audiola — everything in one fileSING-ALONG
Audiola turns any song into your vocal booth: lyrics roll by, the note band shows the original melody, and you see live how well you hit it.
MASTERING
Your mix runs through EQ, compressor and loudness normalization to exactly the LUFS you want. A/B switch for the honest comparison, waveform with click-to-seek.
EQUALIZER & SPATIAL
The interactive 4-band EQ shapes the studio mix live — drag the dots, listen, done. And when stereo isn't enough: place every track freely in space.
VOICES
Create voice profiles from a short recording — locally on your GPU or via ElevenLabs. Swap a clip's voice, generate speech tracks from text.
LIGHT & DARK
Night shift in the dark studio or daylight by the window — drag the handle.
GALLERY
FROM THE MAKERS OF AUDIOLA
The karaoke stage for your party. Drop a song, hand out the mics, sing —
lyrics and notes appear on their own.
Everyone with their own microphone and color — pitch lines dance live across the note band.
From tags, .lrc, Audiola projects — or AI recognition. Each song only once: the song memory remembers everything.
Points for hitting the original melody, combos for streaks. Like back in the day — on your own PC.
Whole playlists or random picks, multiple rounds, standings — until a winner is crowned.
The complete studio: stems, timeline, sing-along, mastering, spatial, voices.
⭳ Get Audiola setupThe karaoke stage: multiplayer, note band, playlists, party.
⭳ Get Singola setupAll versions & changes on GitHub Releases. AI stem separation runs locally via Python + Demucs — the setup wizard helps on first launch.
USED BY
BUILT WITH
Audiola uses components from these open-source projects.
AuralizerAudio visualization
MudBlazor.ExtensionsUI & preview components