LALAL.AI has expanded its VST plugin to support six-stem separation – vocals, bass, drums, acoustic guitar, electric guitar and piano – running entirely offline on the user's local machine, with no internet connection required.
Stem separation inside a DAW has always come with a catch: you break your session to upload audio, wait for processing, download the result and import it back in. LALAL.AI’s expanded VST plugin cuts that loop entirely. Six-stem separation now runs locally, directly inside your session, with no cloud dependency and no usage caps.
Catch up on all the latest news here.
The original plugin covered vocal and instrumental splits. Now, six discrete stems are available: vocals and instrumentals, bass, drums, acoustic guitar, electric guitar and piano. Separating acoustic and electric guitar into their own clean outputs rather than lumping them into a “guitar” or “other” category is a detail that makes a real difference for remixers, post-production engineers and sound designers. The more specific the stem, the less manual work follows.
Separation through the VST is also unlimited. No LALAL.AI credits are used when working within the plugin, so users can extract stems from full-length songs and albums without eating into their account balance. For anyone running repeated iterations on a session or working through long-form material, that adds up quickly.
The plugin runs on Lyra, a model built specifically for local deployment. It’s designed to run efficiently on standard production hardware, not a dedicated server or high-end workstation. LALAL.AI positions Lyra as a tool for iteration rather than final output, prioritising speed and accessibility over maximum fidelity. Separation quality is described as on par with the platform’s previous flagship generations: clean, artifact-reduced stems that hold up in professional contexts without requiring manual cleanup.
“Local stem separation is the future of audio production. With the launch of our six-stem VST, we are proving that the calibre of our algorithms can now live entirely on the user’s machine. We specifically engineered the Lyra model to provide professionals with the perfect equilibrium between processing speed and isolation precision,” says Nik Pogorsky, Product Owner and Co-Founder of LALAL.AI.
There’s a confidentiality case for local processing. Unreleased material, including demos, pre-release albums and client sessions carry legal and contractual weight. Uploading that audio to a third-party server is a risk most professional studios would rather avoid. Local execution removes it entirely.
The plugin supports VST3 on Windows, macOS and Linux, with AU in beta. It’s been tested across Ableton Live, Reaper and FL Studio, along with other VST3-compatible hosts, and is available now to LALAL.AI premium subscribers.
