Download

Get Lattis

The desktop app bundles the daemon and a prebuilt llama-server, so local inference works out of the box.

Coming soon

Prebuilt binaries aren’t published yet. In the meantime, build from source — it’s a single Cargo workspace.

macOS

Apple Silicon · Metal-accelerated local inference; optional MLX.

Linux

x86_64 / arm64 · systemd user service for launch-on-login.

Windows

x86_64 · Scheduled Task for launch-on-login.

From source

Build it yourself

Requires a recent stable Rust toolchain (1.95+), plus curl and tar to fetch the bundled llama-server at build time.

# clone
git clone https://github.com/ZetaMinusOne/lattis.git
cd lattis/app

# build everything (vendors llama-server)
make build

# run the desktop app (it can start the daemon for you)
make run

See the installation guide for MLX setup, launch-on-login, and platform notes.