Private AI,
running locally
Loxa is a local inference studio and CLI for running open models on your own machine. Fast to start, simple to live in — an open alternative to Ollama and LM Studio.
$ loxa run llama3 ↳ pulling llama3 · loading weights · ready ✓ serving on http://localhost:11434 $ loxa pull gemma $ loxa studio # open the desktop UI $
// why loxa
Everything stays on your machine.
No accounts, no cloud round-trips, no metered tokens. Your prompts and your models never leave the hardware you own.
🔒
Private by default
Models run fully offline. Nothing is sent anywhere — your data is yours.
⌘
One simple CLI
Just run, pull, serve, studio. Familiar in seconds if you've used Ollama.
🖥
Studio + terminal
A clean desktop app when you want one, a fast CLI when you don't.
⚡
Native & light
Built in Rust with Tauri. Small footprint, quick cold starts, cross-platform.
// install
Pick your package manager.
Names are reserved across registries while the first release lands.
npm
$ npm i -g @loxadev/cli
cargo
$ cargo install loxa
pip
$ pip install loxa