Reach the panel from another machine

The panel and the engine both bind 127.0.0.1 and neither has any authentication. The panel starts processes and writes files; the engine accepts any bearer token. Do not bind either to a LAN address.

Use an SSH tunnel:

ssh -L 8080:127.0.0.1:8080 -L 1919:127.0.0.1:1919 you@your-box

Then http://127.0.0.1:8080 on your laptop reaches the panel, and anything pointed at http://127.0.0.1:1919 reaches the engine.

Serving other machines on purpose

If you genuinely want other hosts to use the engine — a second workstation, a home-lab service — put a reverse proxy in front of it that terminates TLS and checks a token, and keep llama-server itself on loopback. The Authorization header the engine receives is ignored, so anything that can reach port 1919 can use your GPU and read every prompt sent through it.