Skip to content
MusicianStack
LALAL.AI logo

LALAL.AI

LALAL.AI splits a finished mix into vocals, drums, bass, piano, guitars, and other stems.

LALAL.AI screenshot

LALAL.AI splits a finished mix into vocals, drums, bass, piano, guitars, and other stems. It can also remove background music, noise, and echo from speech. You can use it in a browser or through its desktop and mobile apps. Pro subscribers also get a VST3 plug-in and offline processing.

The main splitter can isolate vocals, instrumental, drums, bass, piano, acoustic guitar, electric guitar, synthesizer, strings, and wind instruments. You select a stem type and receive that stem plus the rest of the mix. Separate tools cover lead and backing vocals, voice cleanup, noise removal, and echo reduction. Audio support includes WAV, FLAC, MP3, AIFF, AAC, M4A, and OGG, while common video formats such as MP4, MOV, MKV, and AVI are also accepted. Paid plans support batch processing for up to 20 files.

The web version runs in a modern browser. Native apps cover Windows 10 and 11, macOS 10.15 or later, Ubuntu 22.04 or later, iOS, and Android. Desktop processing normally uses the cloud models. Pro subscribers can switch to Lyra, the local model, which works offline and keeps source audio on the computer. Lyra is faster and private, while the current cloud models are the better choice for dense mixes and final quality.

DAW support needs a closer look. The plug in is VST3 on Windows, macOS, and Linux, with Audio Units support currently in beta. Ableton Live, Reaper, and VST3 enabled Audacity builds have been tested. There is no AAX version, so Pro Tools users need to process files outside the session. The current macOS build is a Universal 2 binary for Intel and Apple Silicon. Activation is tied to your account and requires occasional internet checks even though audio processing itself can run offline.

The free Starter plan is an audition route rather than a working free tier. It gives you a small minute allowance and result previews, but no full downloads. Lite adds downloads, batch processing, and unlimited use of the slower Relaxed Queue. Pro adds the VST3 plug in, local Lyra processing, and API access. Priority minutes reset each month, and stem jobs deduct the file duration once for every selected separation type. That accounting can burn through the faster allowance sooner than expected. One time top ups are available.

Results are quick and often clean enough for remix sketches, practice tracks, sample preparation, and dialogue work. They are not surgical multitracks. Reverb tails, cymbals, distorted guitars, and layered backing vocals can leave watery edges or spill. Quality changes from song to song, so previewing before committing minutes matters. Logic Pro and newer Ableton Live versions provide more convenient built in splitting, while Ultimate Vocal Remover gives patient users a free local alternative. LALAL.AI makes the most sense when you value speed, broad stem choice, and cross platform access more than detailed spectral editing.

Used For

  • Pulling vocals from a finished song for a remix, mashup, or reference edit
  • Creating instrumental and drumless versions for singers, drummers, and music teachers
  • Isolating bass, piano, guitar, synth, string, or wind parts for transcription and practice
  • Preparing rough stems for DJ edits without access to the original session
  • Cleaning dialogue by reducing background music, noise, mic rumble, echo, and reverb
  • Separating lead and backing vocals for arrangement study or vocal production
  • Splitting audio inside Ableton Live, Reaper, Audacity, and other VST3 hosts
  • Processing private material locally with the Pro tier Lyra model

Pros and Cons

LALAL.AI: Pros Cons
Stem choice:
Vocals plus individual
instruments and voice tools.
Artifacts remain:
Dense mixes and reverb can
leave spill or watery edges.
Platform reach:
Web, desktop, mobile,
and VST3 access.
Free plan limits:
Previews are available,
but full downloads are not.
Local processing:
Lyra works offline and
keeps audio on your machine.
Subscription gate:
VST3 and local Lyra
require the Pro tier.
Fast workflow:
Simple upload, preview,
and batch processing.
No AAX:
Pro Tools needs an
external processing step.
Cloud quality:
Multiple models suit
different source material.
Minute accounting:
Each selected stem type
uses the file duration again.

Also in Audio Repair & Restoration

View all →
001

Audacity

Free
002

iZotope RX 11

Paid