Install lo0

lo0 is one binary with nothing to install beside it. It runs on macOS, Linux and Windows, on amd64 and arm64, and detects Apple Silicon, NVIDIA and AMD GPUs.

macOS and Linux

curl -fsSL https://lo0.ai/install.sh | sh

The script picks the build for your system, downloads it from the GitHub releases, checks it against the release’s checksums.txt and installs it to ~/.local/bin. It changes nothing else. It is a short POSIX shell script, so you can read it before you run it.

Two variables change what it does:

curl -fsSL https://lo0.ai/install.sh | LO0_INSTALL_DIR="$HOME/bin" sh

Windows

In PowerShell:

irm https://lo0.ai/install.ps1 | iex

It installs lo0.exe to %LOCALAPPDATA%\Programs\lo0, checks it against the release checksum and adds the folder to your user PATH. Open a new terminal afterwards. The same LO0_INSTALL_DIR and LO0_VERSION variables apply. Read the script.

With Go

go install lo0.ai/lo0/cmd/lo0@latest

Needs Go 1.27 or newer. The binary lands in $(go env GOPATH)/bin.

From source

git clone https://github.com/daveearley/lo0.ai
cd lo0.ai
make build    # writes bin/lo0

Run it

lo0

In a terminal, lo0 opens the interactive workbench: how much of your work can run on this machine, every model that fits and the next step. lo0 web shows the same in your browser, and lo0 match prints it as a report. Nothing changes on your machine until you confirm a plan.

You do not need a model runtime first. When you try a model, the plan says what is missing: lo0 sets models up in Ollama on its own, and gives guided steps for LM Studio and llama.cpp.

Uninstall

Delete the binary. If you kept a setup with lo0 migrate, undo it first so lo0 can put your files back:

lo0 migrate --undo
rm ~/.local/bin/lo0

Models lo0 downloaded stay in your runtime until you remove them, from the Installed screen or with the runtime’s own command, such as ollama rm.