[Unit]
Description=Local llama.cpp Qwen model server
After=network-online.target graphical-session.target
StartLimitIntervalSec=300
StartLimitBurst=3

[Service]
Type=simple
ExecStart=%h/.local/libexec/hermes-stack-llama-run
Restart=on-failure
RestartSec=30
Nice=10
CPUQuota=1600%
MemoryHigh=40G
MemoryMax=48G
OOMPolicy=stop
TimeoutStopSec=60

[Install]
WantedBy=default.target