Generate images, video, audio, voices and 3D — and train your own models — all from a single desktop app running on hardware you own. Neutral Studio scans your machine and automatically runs the best model your GPUs can handle, tuning itself so it fills the card without ever running out of memory. No cloud, no lock-in, no telemetry.
status:"in active development — closest to launch"
Features
Everything in one suite
Images, video, audio, voices and 3D — generate any modality from a single app. No tool-hopping, no glue scripts.
Your hardware, your data
Runs locally on your own GPUs. Nothing leaves your machine — no cloud middleman, no telemetry, no eavesdropping.
Auto-scales to your machine
Scans your VRAM and memory, picks the best model your rig can run, and tunes it to fill the GPU without ever crashing.
Bring any model
Model-agnostic by design. Drop in any model and Studio detects it, routes it, and runs it — no config required.
Train your own models
A built-in training flywheel turns your work into your own models — shipped in sizes that fit any machine.
Voice-driven
Talk to your studio. Drive the whole pipeline by voice, with your own voice profiles as the assistant.
Open source at the core
Studio's core is open and auditable — no black boxes, no hidden network calls, no telemetry. Read the code, run it yourself, and build on it. Security and trust are the point.