[Unit]
Description=Local LLM Proxy
Wants=basic.target
After=basic.target network.target

[Service]
User=server
Group=server
WorkingDirectory=/srv/server/local-llm-server
# Need a lot of workers since we have long-running requests
# Takes about 3.5G memory
ExecStart=/srv/server/local-llm-server/venv/bin/gunicorn --workers 20 --bind 0.0.0.0:5000 server:app --timeout 60 --worker-class gevent
Restart=always
RestartSec=2

[Install]
WantedBy=multi-user.target