Installation

To install from the official release:

curl -fsSL https://pkg.gpud.dev/install.sh | sh

Running GPUd

• Run with Lepton AI Platform:

Please ensure that your machine has a public IP address and that the GPUd port (default 15132) is reachable.

1. Sign up at lepton.ai and get the token from the left sidebar Tokens in the Settings Tab.
2. Run the gpud command with the token (note: token should be of format workspace:token) :

sudo syntax up --token <TOKEN>

• Run locally with systemd:

sudo syntax up

To connect a locally running gpud to Lepton AI Platform:

sudo syntax login --token <TOKEN>

If your system doesn't have systemd

nohup sudo /usr/sbin/syntax run &>> <your log file path> &

Stop and uninstall

sudo syntax down
sudo rm /usr/sbin/syntax
sudo rm /etc/systemd/system/syntax.service

Key Features

• Monitor critical GPU and GPU fabric metrics (power, temperature).
• Reports GPU and GPU fabric status (nvidia-smi parser, error checking).
• Detects critical GPU and GPU fabric errors (dmesg, hardware slowdown, NVML Xid event, DCGM).
• Monitor overall system metrics (CPU, memory, disk).

Check out components for a detailed list of components and their features.